https://korshunov.ai/en/article/18228-vita-clinical-rag-matches-or-outperforms-frontier-llms-on-healthbench/korshunov.aien2026-08-13T05:35:02.604353+00:00VITA clinical RAG matches or outperforms frontier LLMs on HealthBench https://korshunov.ai/ru/article/18228-vita-klinicheskii-rag-sootvetstvuet-ili-prevoskhodit-peredovye-llm-v-healthbench/korshunov.airu2026-08-13T05:35:02.604353+00:00VITA клинический RAG соответствует или превосходит передовые LLM в HealthBench https://korshunov.ai/es/article/18228-vita-clinico-rag-iguala-o-supera-a-los-llm-de-vanguardia-en-healthbench/korshunov.aies2026-08-13T05:35:02.604353+00:00VITA clínico RAG iguala o supera a los LLM de vanguardia en HealthBench https://korshunov.ai/zh-hans/article/18228-vita-lin-chuang-rag-zai-healthbench-shang-pi-pei-huo-chao-yue-qian-yan-da-yu-yan/korshunov.aizh-cn2026-08-13T05:35:02.604353+00:00VITA 临床 RAG 在 HealthBench 上匹配或超越前沿大语言模型 https://korshunov.ai/pt-br/article/18228-rag-clinico-vita-iguala-ou-supera-llms-de-ponta-no-healthbench/korshunov.aipt2026-08-13T05:35:02.604353+00:00RAG clínico VITA iguala ou supera LLMs de ponta no HealthBench https://korshunov.ai/hi/article/18228-vita-klinikl-rag-healthbench-pr-phrnttiyr-llms-ke-braabr-yaa-behtr-prdrshn-krtaa/korshunov.aihi2026-08-13T05:35:02.604353+00:00VITA क्लिनिकल RAG HealthBench पर फ्रंटियर LLMs के बराबर या बेहतर प्रदर्शन करता है https://korshunov.ai/ja/article/18228-vitanolin-chuang-yong-raghahealthbenchnioitezui-xian-duan-llmtotong-deng-shang/korshunov.aija2026-08-13T05:35:02.604353+00:00VITAの臨床用RAGはHealthBenchにおいて最先端LLMと同等かそれ以上の性能を示す https://korshunov.ai/ko/article/18228-vita-imsang-ragneun-healthbencheseo-coeceomdan-llmgwa-dongdeunghageona-ireul/korshunov.aiko2026-08-13T05:35:02.604353+00:00VITA 임상 RAG는 HealthBench에서 최첨단 LLM과 동등하거나 이를 능가 https://korshunov.ai/fr/article/18228-le-rag-clinique-vita-egale-ou-surpasse-les-grands-modeles-de-langage-de-pointe/korshunov.aifr2026-08-13T05:35:02.604353+00:00Le RAG clinique VITA égale ou surpasse les grands modèles de langage de pointe sur HealthBench https://korshunov.ai/id/article/18228-rag-klinis-vita-menyamai-atau-mengungguli-llm-terdepan-pada-healthbench/korshunov.aiid2026-08-13T05:35:02.604353+00:00RAG klinis VITA menyamai atau mengungguli LLM terdepan pada HealthBench https://korshunov.ai/ar/article/18228-yttbq-vita-lsryry-lm-zz-blstrj-w-ytfwq-l-nmdhj-llgt-lkbyr-lmtqdm-fy-healthbench/korshunov.aiar2026-08-13T05:35:02.604353+00:00يتطابق VITA السريري المعزز بالاسترجاع أو يتفوق على نماذج اللغات الكبيرة المتقدمة في HealthBench https://korshunov.ai/en/article/18096-xai-releases-grok-4-6-alibaba-drops-qwen3-8-max-deepseek-v4-pro-ga/korshunov.aien2026-08-13T04:34:55.913572+00:00xAI releases Grok 4.6, Alibaba drops Qwen3.8-Max, DeepSeek V4 Pro GA https://korshunov.ai/ru/article/18096-xai-vypustila-grok-4-6-alibaba-predstavila-qwen3-8-max-deepseek-v4-pro-stala/korshunov.airu2026-08-13T04:34:55.913572+00:00xAI выпустила Grok 4.6, Alibaba представила Qwen3.8-Max, DeepSeek V4 Pro стала общедоступной https://korshunov.ai/es/article/18096-xai-lanza-grok-4-6-alibaba-suelta-qwen3-8-max-deepseek-v4-pro-ga/korshunov.aies2026-08-13T04:34:55.913572+00:00xAI lanza Grok 4.6, Alibaba suelta Qwen3.8-Max, DeepSeek V4 Pro GA https://korshunov.ai/zh-hans/article/18096-xaifa-bu-grok-4-6-a-li-ba-ba-tui-chu-qwen3-8-max-deepseek-v4-prozheng-shi-ga/korshunov.aizh-cn2026-08-13T04:34:55.913572+00:00xAI发布Grok 4.6，阿里巴巴推出Qwen3.8-Max，DeepSeek V4 Pro正式GA https://korshunov.ai/pt-br/article/18096-xai-lanca-grok-4-6-alibaba-lanca-qwen3-8-max-deepseek-v4-pro-ga/korshunov.aipt2026-08-13T04:34:55.913572+00:00xAI lança Grok 4.6, Alibaba lança Qwen3.8-Max, DeepSeek V4 Pro GA https://korshunov.ai/hi/article/18096-xai-ne-grok-4-6-jaarii-kiyaa-alibaba-ne-qwen3-8-max-choddaa-deepseek-v4-pro-ga/korshunov.aihi2026-08-13T04:34:55.913572+00:00xAI ने Grok 4.6 जारी किया, Alibaba ने Qwen3.8-Max छोड़ा, DeepSeek V4 Pro GA https://korshunov.ai/ja/article/18096-xaikagrok-4-6woriri-su-alibabakaqwen3-8-maxwogong-kai-deepseek-v4-prokagahe/korshunov.aija2026-08-13T04:34:55.913572+00:00xAIがGrok 4.6をリリース、AlibabaがQwen3.8-Maxを公開、DeepSeek V4 ProがGAへ https://korshunov.ai/ko/article/18096-xaiga-grok-4-6-culsi-alribabaga-qwen3-8-max-gonggae-deepseek-v4-pro-ga-balpyo/korshunov.aiko2026-08-13T04:34:55.913572+00:00xAI가 Grok 4.6 출시, 알리바바가 Qwen3.8-Max 공개, DeepSeek V4 Pro GA 발표 https://korshunov.ai/fr/article/18096-xai-publie-grok-4-6-alibaba-lance-qwen3-8-max-deepseek-v4-pro-en-ga/korshunov.aifr2026-08-13T04:34:55.913572+00:00xAI publie Grok 4.6, Alibaba lance Qwen3.8-Max, DeepSeek V4 Pro en GA https://korshunov.ai/id/article/18096-xai-merilis-grok-4-6-alibaba-melepas-qwen3-8-max-deepseek-v4-pro-ga/korshunov.aiid2026-08-13T04:34:55.913572+00:00xAI merilis Grok 4.6, Alibaba melepas Qwen3.8-Max, DeepSeek V4 Pro GA https://korshunov.ai/ar/article/18096-xai-ttlq-grok-4-6-wlyb-ttrh-qwen3-8-max-wdeepseek-v4-pro-mth-mwmyan/korshunov.aiar2026-08-13T04:34:55.913572+00:00xAI تطلق Grok 4.6، وأليبا تطرح Qwen3.8-Max، وDeepSeek V4 Pro متاحة عموميًا https://korshunov.ai/en/article/18157-corpus-specific-vita-rag-matches-or-beats-frontier-llms-on-healthbench/korshunov.aien2026-08-13T03:17:03.269508+00:00Corpus-specific VITA RAG matches or beats frontier LLMs on HealthBench https://korshunov.ai/ru/article/18157-spetsializirovannaia-vita-rag-sootvetstvuet-ili-prevoskhodit-peredovye-llm-v/korshunov.airu2026-08-13T03:17:03.269508+00:00Специализированная VITA RAG соответствует или превосходит передовые LLM в HealthBench https://korshunov.ai/es/article/18157-vita-especifico-del-corpus-iguala-o-supera-a-los-llm-de-vanguardia-en/korshunov.aies2026-08-13T03:17:03.269508+00:00VITA específico del corpus iguala o supera a los LLM de vanguardia en HealthBench https://korshunov.ai/zh-hans/article/18157-zhen-dui-te-ding-yu-liao-de-vita-ragzai-healthbenchshang-pi-pei-huo-chao-yue-yan/korshunov.aizh-cn2026-08-13T03:17:03.269508+00:00针对特定语料的VITA RAG在HealthBench上匹配或超越前沿LLM https://korshunov.ai/pt-br/article/18157-vita-rag-especifico-do-corpus-iguala-ou-supera-llms-de-ponta-no-healthbench/korshunov.aipt2026-08-13T03:17:03.269508+00:00VITA RAG específico do corpus iguala ou supera LLMs de ponta no HealthBench https://korshunov.ai/hi/article/18157-healthbench-pr-corpus-vishisstt-vita-rag-phrnttiyr-llms-ke-braabr-yaa-behtr/korshunov.aihi2026-08-13T03:17:03.269508+00:00HealthBench पर Corpus-विशिष्ट VITA RAG फ्रंटियर LLMs के बराबर या बेहतर https://korshunov.ai/ja/article/18157-corpusgu-you-novita-raghahealthbenchtezui-xian-duan-llmnipi-di-matahashang-hui/korshunov.aija2026-08-13T03:17:03.269508+00:00Corpus固有のVITA RAGはHealthBenchで最先端LLMに匹敵または上回る https://korshunov.ai/ko/article/18157-corpus-teughwa-vita-ragga-healthbencheseo-coeceomdan-llmeul-cugyeoghageona/korshunov.aiko2026-08-13T03:17:03.269508+00:00Corpus 특화 VITA RAG가 HealthBench에서 최첨단 LLM을 추격하거나 능가 https://korshunov.ai/fr/article/18157-le-vita-rag-specifique-au-corpus-egale-ou-bat-les-llm-de-pointe-sur-healthbench/korshunov.aifr2026-08-13T03:17:03.269508+00:00Le VITA RAG spécifique au corpus égale ou bat les LLM de pointe sur HealthBench https://korshunov.ai/id/article/18157-vita-rag-khusus-korpus-cocok-atau-mengungguli-llm-mutakhir-di-healthbench/korshunov.aiid2026-08-13T03:17:03.269508+00:00VITA RAG khusus korpus cocok atau mengungguli LLM mutakhir di HealthBench https://korshunov.ai/ar/article/18157-yttbq-vita-lkhs-bmjmw-lbynt-w-ytfwq-l-nmdhj-llgt-lkbyr-lmtqdm-fy-healthbench/korshunov.aiar2026-08-13T03:17:03.269508+00:00يتطابق VITA الخاص بمجموعة البيانات أو يتفوق على نماذج اللغات الكبيرة المتقدمة في HealthBench https://korshunov.ai/zh-hans/article/18112-vitalin-chuang-ragxi-tong-zai-healthbenchshang-chi-ping-huo-chao-yue-qian-yan/korshunov.aizh-cn2026-08-13T02:48:40.496885+00:00VITA临床RAG系统在HealthBench上持平或超越前沿LLM https://korshunov.ai/pt-br/article/18112-sistema-rag-clinico-vita-iguala-ou-supera-llms-de-ponta-no-healthbench/korshunov.aipt2026-08-13T02:48:40.496885+00:00Sistema RAG clínico VITA iguala ou supera LLMs de ponta no HealthBench https://korshunov.ai/hi/article/18112-healthbench-pr-vita-klinikl-rag-sisttm-phrnttiyr-llms-ke-braabr-yaa-behtr-hai/korshunov.aihi2026-08-13T02:48:40.496885+00:00HealthBench पर VITA क्लिनिकल RAG सिस्टम फ्रंटियर LLMs के बराबर या बेहतर है https://korshunov.ai/ja/article/18112-vitalin-chuang-ragshisutemuhahealthbenchtezui-xian-duan-llmtotong-deng-shang/korshunov.aija2026-08-13T02:48:40.496885+00:00VITA臨床RAGシステムはHealthBenchで最先端LLMと同等またはそれ以上の性能 https://korshunov.ai/ko/article/18112-vita-imsang-rag-siseutemi-healthbencheseo-coejeonbang-llmgwa-dongdeunghageona/korshunov.aiko2026-08-13T02:48:40.496885+00:00VITA 임상 RAG 시스템이 HealthBench에서 최전방 LLM과 동등하거나 우수함 https://korshunov.ai/fr/article/18112-le-systeme-rag-clinique-vita-egale-ou-surpasse-les-llm-de-pointe-sur-healthbench/korshunov.aifr2026-08-13T02:48:40.496885+00:00Le système RAG clinique VITA égale ou surpasse les LLM de pointe sur HealthBench https://korshunov.ai/id/article/18112-sistem-rag-klinis-vita-menyamai-atau-mengungguli-llm-terdepan-di-healthbench/korshunov.aiid2026-08-13T02:48:40.496885+00:00Sistem RAG klinis VITA menyamai atau mengungguli LLM terdepan di HealthBench https://korshunov.ai/ar/article/18112-nzm-vita-llrag-lsryry-yswy-w-ytfwq-l-nmdhj-llm-lmtqdm-fy-healthbench/korshunov.aiar2026-08-13T02:48:40.496885+00:00نظام VITA للرAG السريري يساوي أو يتفوق على نماذج LLM المتقدمة في HealthBench https://korshunov.ai/en/article/18112-vita-clinical-rag-system-matches-or-outperforms-frontier-llms-on-healthbench/korshunov.aien2026-08-13T02:48:40.496885+00:00VITA clinical RAG system matches or outperforms frontier LLMs on HealthBench https://korshunov.ai/ru/article/18112-klinicheskaia-rag-sistema-vita-sootvetstvuet-ili-prevoskhodit-peredovye-llm-v/korshunov.airu2026-08-13T02:48:40.496885+00:00Клиническая RAG-система VITA соответствует или превосходит передовые LLM в HealthBench https://korshunov.ai/es/article/18112-el-sistema-rag-clinico-vita-iguala-o-supera-a-los-llm-de-vanguardia-en/korshunov.aies2026-08-13T02:48:40.496885+00:00El sistema RAG clínico VITA iguala o supera a los LLM de vanguardia en HealthBench https://korshunov.ai/en/article/18095-xai-releases-grok-4-6-alibaba-opens-qwen3-8-moe-deepseek-v4-pro-ga/korshunov.aien2026-08-13T02:20:50.701804+00:00xAI releases Grok 4.6; Alibaba opens Qwen3.8-MoE; DeepSeek V4 Pro GA https://korshunov.ai/ru/article/18095-xai-vypustila-grok-4-6-alibaba-otkryla-qwen3-8-moe-deepseek-v4-pro-ga/korshunov.airu2026-08-13T02:20:50.701804+00:00xAI выпустила Grok 4.6; Alibaba открыла Qwen3.8-MoE; DeepSeek V4 Pro GA https://korshunov.ai/es/article/18095-xai-lanza-grok-4-6-alibaba-abre-qwen3-8-moe-deepseek-v4-pro-ga/korshunov.aies2026-08-13T02:20:50.701804+00:00xAI lanza Grok 4.6; Alibaba abre Qwen3.8-MoE; DeepSeek V4 Pro GA https://korshunov.ai/zh-hans/article/18095-xaifa-bu-grok-4-6-a-li-ba-ba-kai-yuan-qwen3-8-moe-deepseek-v4-prozheng-shi-shang/korshunov.aizh-cn2026-08-13T02:20:50.701804+00:00xAI发布Grok 4.6；阿里巴巴开源Qwen3.8-MoE；DeepSeek V4 Pro正式商用 https://korshunov.ai/pt-br/article/18095-xai-lanca-grok-4-6-alibaba-abre-qwen3-8-moe-deepseek-v4-pro-ga/korshunov.aipt2026-08-13T02:20:50.701804+00:00xAI lança Grok 4.6; Alibaba abre Qwen3.8-MoE; DeepSeek V4 Pro GA https://korshunov.ai/hi/article/18095-xai-ne-grok-4-6-jaarii-kiyaa-alibaba-ne-qwen3-8-moe-ko-kholaa-deepseek-v4-pro-ga/korshunov.aihi2026-08-13T02:20:50.701804+00:00xAI ने Grok 4.6 जारी किया; Alibaba ने Qwen3.8-MoE को खोला; DeepSeek V4 Pro GA https://korshunov.ai/ja/article/18095-xaikagrok-4-6woriri-su-alibabakaqwen3-8-moewogong-kai-deepseek-v4-pro-ga/korshunov.aija2026-08-13T02:20:50.701804+00:00xAIがGrok 4.6をリリース；AlibabaがQwen3.8-MoEを公開；DeepSeek V4 Pro GA https://korshunov.ai/ko/article/18095-xai-grok-4-6-culsi-alribaba-qwen3-8-moe-opeun-deepseek-v4-pro-ga/korshunov.aiko2026-08-13T02:20:50.701804+00:00xAI, Grok 4.6 출시; 알리바바, Qwen3.8-MoE 오픈; DeepSeek V4 Pro GA https://korshunov.ai/fr/article/18095-xai-publie-grok-4-6-alibaba-ouvre-qwen3-8-moe-deepseek-v4-pro-ga/korshunov.aifr2026-08-13T02:20:50.701804+00:00xAI publie Grok 4.6 ; Alibaba ouvre Qwen3.8-MoE ; DeepSeek V4 Pro GA https://korshunov.ai/id/article/18095-xai-merilis-grok-4-6-alibaba-membuka-qwen3-8-moe-deepseek-v4-pro-ga/korshunov.aiid2026-08-13T02:20:50.701804+00:00xAI merilis Grok 4.6; Alibaba membuka Qwen3.8-MoE; DeepSeek V4 Pro GA https://korshunov.ai/ar/article/18095-xai-ttlq-grok-4-6-lybbaa-tfth-qwen3-8-moe-deepseek-v4-ga/korshunov.aiar2026-08-13T02:20:50.701804+00:00xAI تطلق Grok 4.6؛ أليباबा تفتح Qwen3.8-MoE؛ DeepSeek V4 GA https://korshunov.ai/en/article/18092-deepseek-releases-v4-pro-0813-via-api-on-openrouter/korshunov.aien2026-08-13T00:37:14.156416+00:00DeepSeek releases V4 Pro 0813 via API on OpenRouter https://korshunov.ai/ru/article/18092-deepseek-vypustila-v4-pro-0813-cherez-api-na-openrouter/korshunov.airu2026-08-13T00:37:14.156416+00:00DeepSeek выпустила V4 Pro 0813 через API на OpenRouter https://korshunov.ai/es/article/18092-deepseek-lanza-v4-pro-0813-via-api-en-openrouter/korshunov.aies2026-08-13T00:37:14.156416+00:00DeepSeek lanza V4 Pro 0813 vía API en OpenRouter https://korshunov.ai/zh-hans/article/18092-deepseek-tong-guo-openrouter-api-fa-bu-v4-pro-0813/korshunov.aizh-cn2026-08-13T00:37:14.156416+00:00DeepSeek 通过 OpenRouter API 发布 V4 Pro 0813 https://korshunov.ai/pt-br/article/18092-deepseek-lanca-v4-pro-0813-via-api-no-openrouter/korshunov.aipt2026-08-13T00:37:14.156416+00:00DeepSeek lança V4 Pro 0813 via API no OpenRouter https://korshunov.ai/hi/article/18092-deepseek-ne-openrouter-pr-api-ke-maadhym-se-v4-pro-0813-jaarii-kiyaa/korshunov.aihi2026-08-13T00:37:14.156416+00:00DeepSeek ने OpenRouter पर API के माध्यम से V4 Pro 0813 जारी किया https://korshunov.ai/ja/article/18092-deepseekkaopenrouternoapijing-you-tev4-pro-0813woriri-su/korshunov.aija2026-08-13T00:37:14.156416+00:00DeepSeekがOpenRouterのAPI経由でV4 Pro 0813をリリース https://korshunov.ai/ko/article/18092-deepseek-openrouter-api-tonghae-v4-pro-0813-culsi/korshunov.aiko2026-08-13T00:37:14.156416+00:00DeepSeek, OpenRouter API 통해 V4 Pro 0813 출시 https://korshunov.ai/fr/article/18092-deepseek-publie-v4-pro-0813-via-api-sur-openrouter/korshunov.aifr2026-08-13T00:37:14.156416+00:00DeepSeek publie V4 Pro 0813 via API sur OpenRouter https://korshunov.ai/id/article/18092-deepseek-merilis-v4-pro-0813-melalui-api-di-openrouter/korshunov.aiid2026-08-13T00:37:14.156416+00:00DeepSeek merilis V4 Pro 0813 melalui API di OpenRouter https://korshunov.ai/ar/article/18092-dyb-syk-ttlq-v4-pro-0813-br-wjh-brmj-lttbyqt-l-openrouter/korshunov.aiar2026-08-13T00:37:14.156416+00:00ديب سيك تطلق V4 Pro 0813 عبر واجهة برمجة التطبيقات على OpenRouter https://korshunov.ai/en/article/18069-openai-agents-accidentally-attacked-hugging-face-via-artifactory/korshunov.aien2026-08-13T00:12:12.045409+00:00OpenAI agents accidentally attacked Hugging Face via Artifactory https://korshunov.ai/ru/article/18069-agenty-openai-sluchaino-atakovali-hugging-face-cherez-artifactory/korshunov.airu2026-08-13T00:12:12.045409+00:00Агенты OpenAI случайно атаковали Hugging Face через Artifactory https://korshunov.ai/es/article/18069-los-agentes-de-openai-atacaron-accidentalmente-a-hugging-face-a-traves-de/korshunov.aies2026-08-13T00:12:12.045409+00:00Los agentes de OpenAI atacaron accidentalmente a Hugging Face a través de Artifactory https://korshunov.ai/zh-hans/article/18069-openaidai-li-yi-wai-tong-guo-artifactorygong-ji-hugging-face/korshunov.aizh-cn2026-08-13T00:12:12.045409+00:00OpenAI代理意外通过Artifactory攻击Hugging Face https://korshunov.ai/pt-br/article/18069-agentes-da-openai-atacaram-acidentalmente-o-hugging-face-via-artifactory/korshunov.aipt2026-08-13T00:12:12.045409+00:00Agentes da OpenAI atacaram acidentalmente o Hugging Face via Artifactory https://korshunov.ai/hi/article/18069-openai-ke-ejentts-ne-gltii-se-artifactory-ke-maadhym-se-hugging-face-pr-hmlaa/korshunov.aihi2026-08-13T00:12:12.045409+00:00OpenAI के एजेंट्स ने गलती से Artifactory के माध्यम से Hugging Face पर हमला किया https://korshunov.ai/ja/article/18069-openainoe-shientokaartifactorywojie-shitehugging-faceniwu-tsutegong-ji/korshunov.aija2026-08-13T00:12:12.045409+00:00OpenAIのエージェントがArtifactoryを介してHugging Faceに誤って攻撃 https://korshunov.ai/ko/article/18069-openai-eijeonteuga-uyeonhi-artifactoryreul-tonghae-hugging-facereul/korshunov.aiko2026-08-13T00:12:12.045409+00:00OpenAI 에이전트가 우연히 Artifactory를 통해 Hugging Face를 공격하다 https://korshunov.ai/fr/article/18069-les-agents-d-openai-ont-accidentellement-attaque-hugging-face-via-artifactory/korshunov.aifr2026-08-13T00:12:12.045409+00:00Les agents d'OpenAI ont accidentellement attaqué Hugging Face via Artifactory https://korshunov.ai/id/article/18069-agen-openai-secara-tidak-sengaja-menyerang-hugging-face-melalui-artifactory/korshunov.aiid2026-08-13T00:12:12.045409+00:00Agen OpenAI secara tidak sengaja menyerang Hugging Face melalui Artifactory https://korshunov.ai/ar/article/18069-hjmt-wkl-openai-mwq-hugging-face-n-tryq-lkht-br-artifactory/korshunov.aiar2026-08-13T00:12:12.045409+00:00هاجمت وكلاء OpenAI موقع Hugging Face عن طريق الخطأ عبر Artifactory https://korshunov.ai/pt-br/article/18066-anthropic-torna-o-modo-auto-o-padrao-no-claude-code-para-planos-pro-max-e-team/korshunov.aipt2026-08-12T23:48:32.233972+00:00Anthropic torna o modo Auto o padrão no Claude Code para planos Pro, Max e Team https://korshunov.ai/hi/article/18066-anthropic-ne-claude-code-men-pro-max-aur-team-plaanon-ke-lie-auto-modd-ko-bnaa/korshunov.aihi2026-08-12T23:48:32.233972+00:00Anthropic ने Claude Code में Pro, Max और Team प्लानों के लिए Auto मोड को डिफ़ॉल्ट बना दिया है https://korshunov.ai/en/article/18066-anthropic-makes-auto-mode-the-default-in-claude-code-for-pro-max-and-team-plans/korshunov.aien2026-08-12T23:48:32.233972+00:00Anthropic makes Auto mode the default in Claude Code for Pro, Max, and Team plans https://korshunov.ai/ru/article/18066-anthropic-ustanavlivaet-rezhim-auto-po-umolchaniiu-v-claude-code-dlia-tarifov-i/korshunov.airu2026-08-12T23:48:32.233972+00:00Anthropic устанавливает режим Auto по умолчанию в Claude Code для тарифов Pro, Max и Team https://korshunov.ai/es/article/18066-anthropic-establece-el-modo-auto-como-predeterminado-en-claude-code-para-los-pro/korshunov.aies2026-08-12T23:48:32.233972+00:00Anthropic establece el modo Auto como predeterminado en Claude Code para los planes Pro, Max y Team https://korshunov.ai/zh-hans/article/18066-anthropic-jiang-claude-code-de-auto-mo-shi-she-wei-pro-max-he-team-tao-can-de-mo/korshunov.aizh-cn2026-08-12T23:48:32.233972+00:00Anthropic 将 Claude Code 的 Auto 模式设为 Pro、Max 和 Team 套餐的默认设置 https://korshunov.ai/ja/article/18066-anthropic-claude-codenopro-max-teamhuranteautomo-towotehuorutonishe-ding/korshunov.aija2026-08-12T23:48:32.233972+00:00Anthropic、Claude CodeのPro、Max、TeamプランでAutoモードをデフォルトに設定 https://korshunov.ai/ko/article/18066-anthropic-claude-codeyi-pro-max-mic-team-peulraeneseo-auto-modeureul-seoljeong/korshunov.aiko2026-08-12T23:48:32.233972+00:00Anthropic, Claude Code의 Pro, Max 및 Team 플랜에서 Auto 모드를 기본값으로 설정 https://korshunov.ai/fr/article/18066-anthropic-rend-le-mode-auto-par-defaut-dans-claude-code-pour-les-plans-pro-max/korshunov.aifr2026-08-12T23:48:32.233972+00:00Anthropic rend le mode Auto par défaut dans Claude Code pour les plans Pro, Max et Team https://korshunov.ai/id/article/18066-anthropic-menjadikan-mode-auto-sebagai-default-di-claude-code-untuk-paket-pro/korshunov.aiid2026-08-12T23:48:32.233972+00:00Anthropic menjadikan mode Auto sebagai default di Claude Code untuk paket Pro, Max, dan Team https://korshunov.ai/ar/article/18066-anthropic-tj-l-wd-auto-lftrdy-fy-claude-code-lkhtt-pro-w-max-w-team/korshunov.aiar2026-08-12T23:48:32.233972+00:00Anthropic تجعل وضع Auto الافتراضي في Claude Code لخطط Pro و Max و Team https://korshunov.ai/en/article/18061-meta-releases-open-weight-muse-glimmer-30b-agentic-vision-model/korshunov.aien2026-08-12T23:31:49.039773+00:00Meta releases open-weight Muse Glimmer 30B agentic vision model https://korshunov.ai/ru/article/18061-meta-vypustila-otkrytuiu-model-agenta-s-zreniem-muse-glimmer-30b/korshunov.airu2026-08-12T23:31:49.039773+00:00Meta выпустила открытую модель агента с зрением Muse Glimmer 30B https://korshunov.ai/es/article/18061-meta-lanza-el-modelo-de-vision-agentica-muse-glimmer-30b-de-pesos-abiertos/korshunov.aies2026-08-12T23:31:49.039773+00:00Meta lanza el modelo de visión agéntica Muse Glimmer 30B de pesos abiertos https://korshunov.ai/zh-hans/article/18061-metafa-bu-kai-yuan-quan-zhong-muse-glimmer-30bzhi-neng-ti-shi-jue-mo-xing/korshunov.aizh-cn2026-08-12T23:31:49.039773+00:00Meta发布开源权重Muse Glimmer 30B智能体视觉模型 https://korshunov.ai/pt-br/article/18061-meta-lanca-o-modelo-de-visao-agentica-muse-glimmer-30b-de-pesos-abertos/korshunov.aipt2026-08-12T23:31:49.039773+00:00Meta lança o modelo de visão agêntica Muse Glimmer 30B de pesos abertos https://korshunov.ai/hi/article/18061-meta-ne-khule-vjn-vaale-muse-glimmer-30b-ejenttik-vijn-monddl-ko-riliij-kiyaa/korshunov.aihi2026-08-12T23:31:49.039773+00:00Meta ने खुले-वजन वाले Muse Glimmer 30B एजेंटिक विज़न मॉडल को रिलीज़ किया https://korshunov.ai/ja/article/18061-metakao-hunueitonomuse-glimmer-30be-shientohishiyonmoteruworiri-su/korshunov.aija2026-08-12T23:31:49.039773+00:00MetaがオープンウェイトのMuse Glimmer 30Bエージェントビジョンモデルをリリース https://korshunov.ai/ko/article/18061-metaga-opeun-gajungciyi-muse-glimmer-30b-eijeonteu-bijeon-model-culsi/korshunov.aiko2026-08-12T23:31:49.039773+00:00Meta가 오픈 가중치의 Muse Glimmer 30B 에이전트 비전 모델 출시 https://korshunov.ai/fr/article/18061-meta-lance-le-modele-de-vision-agentique-muse-glimmer-30b-a-poids-ouverts/korshunov.aifr2026-08-12T23:31:49.039773+00:00Meta lance le modèle de vision agéntique Muse Glimmer 30B à poids ouverts https://korshunov.ai/id/article/18061-meta-merilis-model-visi-agentic-muse-glimmer-30b-bobot-terbuka/korshunov.aiid2026-08-12T23:31:49.039773+00:00Meta merilis model visi agentic Muse Glimmer 30B bobot terbuka https://korshunov.ai/ar/article/18061-myt-ttlq-nmwdhj-lrwy-lwkyly-muse-glimmer-30b-bwzn-mftwh/korshunov.aiar2026-08-12T23:31:49.039773+00:00ميتا تطلق نموذج الرؤية الوكيلية Muse Glimmer 30B بأوزان مفتوحة https://korshunov.ai/en/article/18059-researchers-steal-reasoning-traces-from-proprietary-llm-apis-via-replay-attacks/korshunov.aien2026-08-12T23:17:41.147204+00:00Researchers steal reasoning traces from proprietary LLM APIs via replay attacks https://korshunov.ai/ru/article/18059-issledovateli-pokhishchaiut-tsepochki-rassuzhdenii-iz-proprietarnykh-api-llm-s/korshunov.airu2026-08-12T23:17:41.147204+00:00Исследователи похищают цепочки рассуждений из проприетарных API LLM с помощью атак повторного воспроизведения https://korshunov.ai/es/article/18059-investigadores-roban-trazos-de-razonamiento-de-apis-de-llm-propietarios-mediante/korshunov.aies2026-08-12T23:17:41.147204+00:00Investigadores roban trazos de razonamiento de APIs de LLM propietarios mediante ataques de repetición https://korshunov.ai/zh-hans/article/18059-yan-jiu-ren-yuan-tong-guo-zhong-fang-gong-ji-cong-zhuan-you-da-yu-yan-mo-xing-qu/korshunov.aizh-cn2026-08-12T23:17:41.147204+00:00研究人员通过重放攻击从专有大语言模型API中窃取推理轨迹 https://korshunov.ai/pt-br/article/18059-pesquisadores-roubam-rastros-de-raciocinio-de-apis-de-llm-proprietarias-por-meio/korshunov.aipt2026-08-12T23:17:41.147204+00:00Pesquisadores roubam rastros de raciocínio de APIs de LLM proprietárias por meio de ataques de replay https://korshunov.ai/hi/article/18059-anusndhaankrtaa-riiple-attaik-ke-maadhym-se-propraaittrii-llm-api-se-trk-ttres/korshunov.aihi2026-08-12T23:17:41.147204+00:00अनुसंधानकर्ता रीप्ले अटैक के माध्यम से प्रोप्राइटरी LLM API से तर्क ट्रेस चुराते हैं https://korshunov.ai/ja/article/18059-yan-jiu-zhe-karihureigong-ji-niyorihurohuraietarinallm-apikaratui-lun-tore-qu/korshunov.aija2026-08-12T23:17:41.147204+00:00研究者がリプレイ攻撃によりプロプライエタリなLLM APIから推論トレースを窃取 https://korshunov.ai/ko/article/18059-yeongujadeuli-ripeulrei-gonggyeogeul-tonghae-dogjeom-llm-apieseo-curon-talcwi/korshunov.aiko2026-08-12T23:17:41.147204+00:00연구자들이 리플레이 공격을 통해 독점 LLM API에서 추론 트레이스를 탈취 https://korshunov.ai/fr/article/18059-des-chercheurs-volent-les-traces-de-raisonnement-des-api-llm-proprietaires-via/korshunov.aifr2026-08-12T23:17:41.147204+00:00Des chercheurs volent les traces de raisonnement des API LLM propriétaires via des attaques par rejeu https://korshunov.ai/id/article/18059-peneliti-mencuri-jejak-penalaran-dari-api-llm-tertutup-melalui-serangan-replay/korshunov.aiid2026-08-12T23:17:41.147204+00:00Peneliti mencuri jejak penalaran dari API LLM tertutup melalui serangan replay https://korshunov.ai/ar/article/18059-bhthwn-ysrqwn-msrt-lstdll-mn-wjht-brmj-lttbyqt-lkhs-lnmdhj-llgt-lkbyr-br-hjmt-d/korshunov.aiar2026-08-12T23:17:41.147204+00:00باحثون يسرقون مسارات الاستدلال من واجهات برمجة التطبيقات الخاصة لنماذج اللغات الكبيرة عبر هجمات إعادة التشغيل https://korshunov.ai/en/article/18049-ai-agents-debate-why-they-avoid-saying-i-don-t-know/korshunov.aien2026-08-12T22:40:03.141719+00:00AI agents debate why they avoid saying "I don't know" https://korshunov.ai/ru/article/18049-ii-agenty-obsuzhdaiut-pochemu-oni-izbegaiut-govorit-ia-ne-znaiu/korshunov.airu2026-08-12T22:40:03.141719+00:00ИИ-агенты обсуждают, почему они избегают говорить «Я не знаю» https://korshunov.ai/es/article/18049-los-agentes-de-ia-debaten-por-que-evitan-decir-no-lo-se/korshunov.aies2026-08-12T22:40:03.141719+00:00Los agentes de IA debaten por qué evitan decir "No lo sé" https://korshunov.ai/pt-br/article/18049-agentes-de-ia-debatem-por-que-evitam-dizer-nao-sei/korshunov.aipt2026-08-12T22:40:03.141719+00:00Agentes de IA debatem por que evitam dizer "Não sei" https://korshunov.ai/fr/article/18049-les-agents-ia-debattent-de-pourquoi-ils-evitent-de-dire-je-ne-sais-pas/korshunov.aifr2026-08-12T22:40:03.141719+00:00Les agents IA débattent de pourquoi ils évitent de dire "Je ne sais pas" https://korshunov.ai/hi/article/18049-ai-ejentts-bhs-krte-hain-ki-ve-mujhe-nhiin-ptaa-kyon-khne-se-bcte-hain/korshunov.aihi2026-08-12T22:40:03.141719+00:00AI एजेंट्स बहस करते हैं कि वे "मुझे नहीं पता" क्यों कहने से बचते हैं https://korshunov.ai/ja/article/18049-aie-shientoka-wakarimasen-toyan-wanaili-you-nitsuiteyi-lun/korshunov.aija2026-08-12T22:40:03.141719+00:00AIエージェントが「わかりません」と言わない理由について議論 https://korshunov.ai/ko/article/18049-ai-eijeonteuga-moreugessseubnida-rago-malhaneun-geoseul-pihaneun-iyue-daehan/korshunov.aiko2026-08-12T22:40:03.141719+00:00AI 에이전트가 "모르겠습니다"라고 말하는 것을 피하는 이유에 대한 토론 https://korshunov.ai/id/article/18049-agen-ai-berdebat-mengapa-mereka-menghindari-mengatakan-saya-tidak-tahu/korshunov.aiid2026-08-12T22:40:03.141719+00:00Agen AI berdebat mengapa mereka menghindari mengatakan "Saya tidak tahu" https://korshunov.ai/ar/article/18049-wkl-ldhk-lstn-y-ynqshwn-sbb-tjnbhm-qwl-l-rf/korshunov.aiar2026-08-12T22:40:03.141719+00:00وكلاء الذكاء الاصطناعي يناقشون سبب تجنبهم قول "لا أعرف" https://korshunov.ai/en/article/18048-author-proposes-open-causal-seal-for-verifiable-ai-decision-provenance/korshunov.aien2026-08-12T22:34:34.040300+00:00Author proposes open 'Causal Seal' for verifiable AI decision provenance https://korshunov.ai/ru/article/18048-avtor-predlagaet-otkrytyi-prichinnyi-pechat-dlia-proveriaemogo-proiskhozhdeniia/korshunov.airu2026-08-12T22:34:34.040300+00:00Автор предлагает открытый «Причинный Печать» для проверяемого происхождения решений ИИ https://korshunov.ai/es/article/18048-el-autor-propone-un-sello-causal-abierto-para-la-procedencia-verificable-de-de/korshunov.aies2026-08-12T22:34:34.040300+00:00El autor propone un "Sello Causal" abierto para la procedencia verificable de decisiones de IA https://korshunov.ai/zh-hans/article/18048-zuo-zhe-ti-chu-kai-fang-de-yin-guo-yin-zhang-yi-yan-zheng-aijue-ce-su-yuan/korshunov.aizh-cn2026-08-12T22:34:34.040300+00:00作者提出开放的“因果印章”以验证AI决策溯源 https://korshunov.ai/pt-br/article/18048-autor-propoe-selos-causais-abertos-para-proveniencia-verificavel-de-decisoes-de/korshunov.aipt2026-08-12T22:34:34.040300+00:00Autor propõe "Selos Causais" abertos para proveniência verificável de decisões de IA https://korshunov.ai/hi/article/18048-lekhk-ai-nirnny-prmaann-ke-lie-khulaa-kaarnn-muhr-prstaavit-krtaa-hai/korshunov.aihi2026-08-12T22:34:34.040300+00:00लेखक AI निर्णय प्रमाण के लिए खुला 'कारण मुहर' प्रस्तावित करता है https://korshunov.ai/ja/article/18048-zhu-zhe-kaaipan-duan-nojian-zheng-ke-neng-nayou-lai-notamenoo-hunna-yin-guo-shi/korshunov.aija2026-08-12T22:34:34.040300+00:00著者がAI判断の検証可能な由来のためのオープンな「因果シール」を提案 https://korshunov.ai/ko/article/18048-jeojaga-geomjeung-ganeunghan-ai-yisagyeoljeong-gyeboreul-wihan-gaebanghyeong/korshunov.aiko2026-08-12T22:34:34.040300+00:00저자가 검증 가능한 AI 의사결정 계보를 위한 개방형 '인과 봉인' 제안 https://korshunov.ai/fr/article/18048-l-auteur-propose-un-sceau-causal-ouvert-pour-la-provenance-verifiable-des-d-ia/korshunov.aifr2026-08-12T22:34:34.040300+00:00L'auteur propose un "Sceau Causal" ouvert pour la provenance vérifiable des décisions d'IA https://korshunov.ai/id/article/18048-penulis-mengusulkan-segel-kausal-terbuka-untuk-asal-usul-keputusan-ai-yang-dapat/korshunov.aiid2026-08-12T22:34:34.040300+00:00Penulis mengusulkan 'Segel Kausal' terbuka untuk asal-usul keputusan AI yang dapat diverifikasi https://korshunov.ai/ar/article/18048-lmwlf-yqtrh-lkhtm-lsbby-lmftwh-lsly-qrrt-ldhk-lstn-y-lqbl-llthqq/korshunov.aiar2026-08-12T22:34:34.040300+00:00المؤلف يقترح "الختم السببي" المفتوح لأصلية قرارات الذكاء الاصطناعي القابلة للتحقق https://korshunov.ai/en/article/18047-microsoft-research-introduces-mindtopo-benchmark-for-evaluating-vlms-spatial/korshunov.aien2026-08-12T22:25:29.866631+00:00Microsoft Research introduces MindTopo benchmark for evaluating VLMs' spatial reasoning https://korshunov.ai/ru/article/18047-microsoft-research-predstavliaet-benchmark-mindtopo-dlia-otsenki-myshleniia-vlm/korshunov.airu2026-08-12T22:25:29.866631+00:00Microsoft Research представляет бенчмарк MindTopo для оценки пространственного мышления VLM https://korshunov.ai/es/article/18047-microsoft-research-presenta-el-benchmark-mindtopo-para-evaluar-el-razonamiento/korshunov.aies2026-08-12T22:25:29.866631+00:00Microsoft Research presenta el benchmark MindTopo para evaluar el razonamiento espacial de los VLM https://korshunov.ai/zh-hans/article/18047-wei-ruan-yan-jiu-yuan-tui-chu-mindtopoji-zhun-yong-yu-ping-gu-vlmde-kong-jian-li/korshunov.aizh-cn2026-08-12T22:25:29.866631+00:00微软研究院推出MindTopo基准，用于评估VLM的空间推理能力 https://korshunov.ai/pt-br/article/18047-microsoft-research-apresenta-benchmark-mindtopo-para-avaliar-o-raciocinio-de/korshunov.aipt2026-08-12T22:25:29.866631+00:00Microsoft Research apresenta benchmark MindTopo para avaliar o raciocínio espacial de VLMs https://korshunov.ai/hi/article/18047-maaikrosonphtt-risrc-ne-vlms-kii-sthaanik-trkshkti-kaa-muulyaankn-krne-ke-lie/korshunov.aihi2026-08-12T22:25:29.866631+00:00माइक्रोसॉफ्ट रिसर्च ने VLMs की स्थानिक तर्कशक्ति का मूल्यांकन करने के लिए MindTopo बेंचमार्क पेश किया https://korshunov.ai/ja/article/18047-microsoft-research-vlmnokong-jian-tui-lun-woping-jia-surutamenomindtopohenchima/korshunov.aija2026-08-12T22:25:29.866631+00:00Microsoft Research、VLMの空間推論を評価するためのMindTopoベンチマークを導入 https://korshunov.ai/ko/article/18047-maikeurosopeuteu-riseoci-vlmyi-gonggan-curon-pyeonggareul-wihan-mindtopo-balpyo/korshunov.aiko2026-08-12T22:25:29.866631+00:00마이크로소프트 리서치, VLM의 공간 추론 평가를 위한 MindTopo 벤치마크 발표 https://korshunov.ai/fr/article/18047-microsoft-research-presente-le-benchmark-mindtopo-pour-evaluer-le-raisonnement/korshunov.aifr2026-08-12T22:25:29.866631+00:00Microsoft Research présente le benchmark MindTopo pour évaluer le raisonnement spatial des VLM https://korshunov.ai/id/article/18047-microsoft-research-memperkenalkan-benchmark-mindtopo-untuk-mengevaluasi-spasial/korshunov.aiid2026-08-12T22:25:29.866631+00:00Microsoft Research memperkenalkan benchmark MindTopo untuk mengevaluasi penalaran spasial VLM https://korshunov.ai/ar/article/18047-bhth-mykrwswft-tqdm-m-yr-mindtopo-ltqyym-lstdll-lmkny-lnmdhj-vlms/korshunov.aiar2026-08-12T22:25:29.866631+00:00أبحاث مايكروسوفت تقدم معيار MindTopo لتقييم الاستدلال المكاني لنماذج VLMs https://korshunov.ai/en/article/18046-claude-code-v2-1-229-adds-remote-control-resumption-and-fixes-crashes/korshunov.aien2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 adds remote control resumption and fixes crashes https://korshunov.ai/ru/article/18046-claude-code-v2-1-229-dobavliaet-vozobnovlenie-udalennogo-upravleniia-i-sboi/korshunov.airu2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 добавляет возобновление удалённого управления и исправляет сбои https://korshunov.ai/es/article/18046-claude-code-v2-1-229-anade-la-reanudacion-del-control-remoto-y-corrige-fallos/korshunov.aies2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 añade la reanudación del control remoto y corrige fallos https://korshunov.ai/zh-hans/article/18046-claude-code-v2-1-229-xin-zeng-yuan-cheng-kong-zhi-hui-fu-gong-neng-bing-xiu-fu/korshunov.aizh-cn2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 新增远程控制恢复功能并修复崩溃问题 https://korshunov.ai/pt-br/article/18046-claude-code-v2-1-229-adiciona-retomada-de-controle-remoto-e-corrige-falhas/korshunov.aipt2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 adiciona retomada de controle remoto e corrige falhas https://korshunov.ai/hi/article/18046-claude-code-v2-1-229-men-rimott-knttrol-riijmpshn-joddaa-gyaa-aur-kraish-tthiik/korshunov.aihi2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 में रिमोट कंट्रोल रीजम्पशन जोड़ा गया और क्रैश ठीक किए गए https://korshunov.ai/ja/article/18046-claude-code-v2-1-229karimo-tozhi-yu-nozai-kai-tokuratsushiyuxiu-zheng-wozhui-jia/korshunov.aija2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229がリモート制御の再開とクラッシュ修正を追加 https://korshunov.ai/ko/article/18046-claude-code-v2-1-229ga-weongyeog-jeeo-jaegae-gineungeul-cugahago-cungdoleul/korshunov.aiko2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229가 원격 제어 재개 기능을 추가하고 충돌을 수정함 https://korshunov.ai/fr/article/18046-claude-code-v2-1-229-ajoute-la-reprise-du-controle-a-distance-et-corrige-des/korshunov.aifr2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 ajoute la reprise du contrôle à distance et corrige des plantages https://korshunov.ai/id/article/18046-claude-code-v2-1-229-menambahkan-kelanjutan-kendali-jarak-jauh-dan-memperbaiki/korshunov.aiid2026-08-12T22:01:42.016815+00:00Claude Code v2.1.229 menambahkan kelanjutan kendali jarak jauh dan memperbaiki kerusakan https://korshunov.ai/ar/article/18046-ydyf-claude-code-v2-1-229-stynf-lthkm-n-bu-d-wyslh-l-tl/korshunov.aiar2026-08-12T22:01:42.016815+00:00يضيف Claude Code v2.1.229 استئناف التحكم عن بُعد ويصلح الأعطال https://korshunov.ai/en/article/18040-xai-releases-grok-4-6-with-improved-long-running-agent-capabilities/korshunov.aien2026-08-12T20:05:41.915433+00:00xAI releases Grok 4.6 with improved long-running agent capabilities https://korshunov.ai/ru/article/18040-xai-vypustila-grok-4-6-s-uluchshennymi-vozmozhnostiami-dolgosrochnykh-agentov/korshunov.airu2026-08-12T20:05:41.915433+00:00xAI выпустила Grok 4.6 с улучшенными возможностями долгосрочных агентов https://korshunov.ai/es/article/18040-xai-lanza-grok-4-6-con-capacidades-mejoradas-para-agentes-de-larga-duracion/korshunov.aies2026-08-12T20:05:41.915433+00:00xAI lanza Grok 4.6 con capacidades mejoradas para agentes de larga duración https://korshunov.ai/zh-hans/article/18040-xaifa-bu-grok-4-6-zeng-qiang-chang-yun-xing-zhi-neng-ti-neng-li/korshunov.aizh-cn2026-08-12T20:05:41.915433+00:00xAI发布Grok 4.6，增强长运行智能体能力 https://korshunov.ai/pt-br/article/18040-xai-lanca-grok-4-6-com-capacidades-aprimoradas-para-agentes-de-longa-duracao/korshunov.aipt2026-08-12T20:05:41.915433+00:00xAI lança Grok 4.6 com capacidades aprimoradas para agentes de longa duração https://korshunov.ai/hi/article/18040-xai-ne-grok-4-6-ko-lonnc-kiyaa-jismen-lnbe-smy-tk-clne-vaale-ejentt-kii-behtr/korshunov.aihi2026-08-12T20:05:41.915433+00:00xAI ने Grok 4.6 को लॉन्च किया, जिसमें लंबे समय तक चलने वाले एजेंट की क्षमताएं बेहतर हैं https://korshunov.ai/ja/article/18040-xaikagrok-4-6woriri-su-chang-shi-jian-shi-xing-e-shientonoji-neng-woqiang-hua/korshunov.aija2026-08-12T20:05:41.915433+00:00xAIがGrok 4.6をリリース、長時間実行エージェントの機能を強化 https://korshunov.ai/ko/article/18040-xai-janggi-silhaeng-eijeonteu-gineung-gaeseondoen-grok-4-6-culsi/korshunov.aiko2026-08-12T20:05:41.915433+00:00xAI, 장기 실행 에이전트 기능 개선된 Grok 4.6 출시 https://korshunov.ai/fr/article/18040-xai-lance-grok-4-6-avec-des-capacites-ameliorees-pour-les-agents-a-longue-duree/korshunov.aifr2026-08-12T20:05:41.915433+00:00xAI lance Grok 4.6 avec des capacités améliorées pour les agents à longue durée https://korshunov.ai/id/article/18040-xai-merilis-grok-4-6-dengan-kemampuan-agen-jangka-panjang-yang-lebih-baik/korshunov.aiid2026-08-12T20:05:41.915433+00:00xAI merilis Grok 4.6 dengan kemampuan agen jangka panjang yang lebih baik https://korshunov.ai/ar/article/18040-xai-ttlq-grok-4-6-bqdrt-mhswn-llwkl-twyly-ltshgyl/korshunov.aiar2026-08-12T20:05:41.915433+00:00xAI تطلق Grok 4.6 بقدرات محسّنة للوكلاء طويلي التشغيل https://korshunov.ai/en/article/18037-alibaba-releases-qwen3-8-2-4t-a95b-open-weights-for-nvidia-gb300-nvl72/korshunov.aien2026-08-12T18:51:25.219642+00:00Alibaba releases Qwen3.8-2.4T-A95B open weights for NVIDIA GB300 NVL72 https://korshunov.ai/ru/article/18037-alibaba-vypustila-otkrytye-vesa-qwen3-8-2-4t-a95b-dlia-nvidia-gb300-nvl72/korshunov.airu2026-08-12T18:51:25.219642+00:00Alibaba выпустила открытые веса Qwen3.8-2.4T-A95B для NVIDIA GB300 NVL72 https://korshunov.ai/es/article/18037-alibaba-libera-los-pesos-abiertos-de-qwen3-8-2-4t-a95b-para-nvidia-gb300-nvl72/korshunov.aies2026-08-12T18:51:25.219642+00:00Alibaba libera los pesos abiertos de Qwen3.8-2.4T-A95B para NVIDIA GB300 NVL72 https://korshunov.ai/zh-hans/article/18037-a-li-ba-ba-fa-bu-gua-yong-yu-nvidia-gb300-nvl72-de-qwen3-8-2-4t-a95b-kai-yuan/korshunov.aizh-cn2026-08-12T18:51:25.219642+00:00阿里巴巴发布适用于 NVIDIA GB300 NVL72 的 Qwen3.8-2.4T-A95B 开源权重 https://korshunov.ai/pt-br/article/18037-alibaba-libera-pesos-abertos-do-qwen3-8-2-4t-a95b-para-nvidia-gb300-nvl72/korshunov.aipt2026-08-12T18:51:25.219642+00:00Alibaba libera pesos abertos do Qwen3.8-2.4T-A95B para NVIDIA GB300 NVL72 https://korshunov.ai/ar/article/18037-lybb-ttlq-lwzn-lmftwh-l-qwen3-8-2-4t-a95b-ljhz-nvidia-gb300-nvl72/korshunov.aiar2026-08-12T18:51:25.219642+00:00أليبابا تطلق الأوزان المفتوحة لـ Qwen3.8-2.4T-A95B لجهاز NVIDIA GB300 NVL72 https://korshunov.ai/hi/article/18037-alibaba-ne-nvidia-gb300-nvl72-ke-lie-qwen3-8-2-4t-a95b-opn-vetts-jaarii-kie/korshunov.aihi2026-08-12T18:51:25.219642+00:00अलibaba ने NVIDIA GB300 NVL72 के लिए Qwen3.8-2.4T-A95B ओपन वेट्स जारी किए https://korshunov.ai/ja/article/18037-arihaha-nvidia-gb300-nvl72xiang-keqwen3-8-2-4t-a95bnoo-hunueitoworiri-su/korshunov.aija2026-08-12T18:51:25.219642+00:00アリババ、NVIDIA GB300 NVL72向けQwen3.8-2.4T-A95Bのオープンウェイトをリリース https://korshunov.ai/ko/article/18037-alribaba-nvidia-gb300-nvl72yong-qwen3-8-2-4t-a95b-opeun-gajungci-culsi/korshunov.aiko2026-08-12T18:51:25.219642+00:00알리바바, NVIDIA GB300 NVL72용 Qwen3.8-2.4T-A95B 오픈 가중치 출시 https://korshunov.ai/fr/article/18037-alibaba-publie-les-poids-ouverts-de-qwen3-8-2-4t-a95b-pour-nvidia-gb300-nvl72/korshunov.aifr2026-08-12T18:51:25.219642+00:00Alibaba publie les poids ouverts de Qwen3.8-2.4T-A95B pour NVIDIA GB300 NVL72 https://korshunov.ai/id/article/18037-alibaba-merilis-bobot-terbuka-qwen3-8-2-4t-a95b-untuk-nvidia-gb300-nvl72/korshunov.aiid2026-08-12T18:51:25.219642+00:00Alibaba merilis bobot terbuka Qwen3.8-2.4T-A95B untuk NVIDIA GB300 NVL72 https://korshunov.ai/zh-hans/article/18029-olmoearth-studio-xin-zeng-zi-ding-yi-qian-ru-dao-chu-gong-neng-zhi-chi-xia-you/korshunov.aizh-cn2026-08-12T16:49:08.346021+00:00OlmoEarth Studio 新增自定义嵌入导出功能，支持下游分析 https://korshunov.ai/en/article/18029-olmoearth-studio-adds-custom-embedding-exports-for-downstream-analysis/korshunov.aien2026-08-12T16:49:08.346021+00:00OlmoEarth Studio adds custom embedding exports for downstream analysis https://korshunov.ai/ru/article/18029-olmoearth-studio-dobavliaet-eksport-polzovatelskikh-embeddingov-dlia-analiza/korshunov.airu2026-08-12T16:49:08.346021+00:00OlmoEarth Studio добавляет экспорт пользовательских эмбеддингов для последующего анализа https://korshunov.ai/es/article/18029-olmoearth-studio-anade-exportaciones-de-incrustaciones-personalizadas-para/korshunov.aies2026-08-12T16:49:08.346021+00:00OlmoEarth Studio añade exportaciones de incrustaciones personalizadas para análisis posteriores https://korshunov.ai/pt-br/article/18029-olmoearth-studio-adiciona-exportacoes-personalizadas-de-embeddings-para-analise/korshunov.aipt2026-08-12T16:49:08.346021+00:00OlmoEarth Studio adiciona exportações personalizadas de embeddings para análise downstream https://korshunov.ai/hi/article/18029-olmoearth-studio-ne-ddaaunsttriim-vishlessnn-ke-lie-ksttm-embedding-niryaat/korshunov.aihi2026-08-12T16:49:08.346021+00:00OlmoEarth Studio ने डाउनस्ट्रीम विश्लेषण के लिए कस्टम एम्बेडिंग निर्यात जोड़े https://korshunov.ai/ja/article/18029-olmoearth-studiokaxia-liu-fen-xi-yong-nokasutamumai-meip-miekusuho-towozhui-jia/korshunov.aija2026-08-12T16:49:08.346021+00:00OlmoEarth Studioが下流分析用のカスタム埋め込みエクスポートを追加 https://korshunov.ai/ko/article/18029-olmoearth-studioga-haryu-bunseogeul-wihan-sayongja-jeongyi-imbeding/korshunov.aiko2026-08-12T16:49:08.346021+00:00OlmoEarth Studio가 하류 분석을 위한 사용자 정의 임베딩 내보내기를 추가했습니다 https://korshunov.ai/fr/article/18029-olmoearth-studio-ajoute-des-exports-d-embeddings-personnalises-pour-l-analyse-en/korshunov.aifr2026-08-12T16:49:08.346021+00:00OlmoEarth Studio ajoute des exports d'embeddings personnalisés pour l'analyse en aval https://korshunov.ai/id/article/18029-olmoearth-studio-menambahkan-ekspor-embedding-kustom-untuk-analisis-turunan/korshunov.aiid2026-08-12T16:49:08.346021+00:00OlmoEarth Studio menambahkan ekspor embedding kustom untuk analisis turunan https://korshunov.ai/ar/article/18029-ydyf-olmoearth-studio-tsdyran-mkhssan-llmtjht-ltdmyny-llthlyl-llhq/korshunov.aiar2026-08-12T16:49:08.346021+00:00يضيف OlmoEarth Studio تصديرًا مخصصًا للمتجهات التضمينية للتحليل اللاحق https://korshunov.ai/en/article/18020-qwen-releases-qwen3-8-24t-model/korshunov.aien2026-08-12T16:15:12.664522+00:00Qwen releases Qwen3.8-24T model https://korshunov.ai/ru/article/18020-qwen-vypustila-model-qwen3-8-24t/korshunov.airu2026-08-12T16:15:12.664522+00:00Qwen выпустила модель Qwen3.8-24T https://korshunov.ai/es/article/18020-qwen-lanza-el-modelo-qwen3-8-24t/korshunov.aies2026-08-12T16:15:12.664522+00:00Qwen lanza el modelo Qwen3.8-24T https://korshunov.ai/pt-br/article/18020-qwen-lanca-o-modelo-qwen3-8-24t/korshunov.aipt2026-08-12T16:15:12.664522+00:00Qwen lança o modelo Qwen3.8-24T https://korshunov.ai/hi/article/18020-qwen-ne-qwen3-8-24t-monddl-jaarii-kiyaa/korshunov.aihi2026-08-12T16:15:12.664522+00:00Qwen ने Qwen3.8-24T मॉडल जारी किया https://korshunov.ai/fr/article/18020-qwen-lance-le-modele-qwen3-8-24t/korshunov.aifr2026-08-12T16:15:12.664522+00:00Qwen lance le modèle Qwen3.8-24T https://korshunov.ai/id/article/18020-qwen-merilis-model-qwen3-8-24t/korshunov.aiid2026-08-12T16:15:12.664522+00:00Qwen merilis model Qwen3.8-24T https://korshunov.ai/ja/article/18023-calibrakarohoteikusuxue-xi-te-tasetsutonowen-ti-wojian-chu/korshunov.aija2026-08-12T16:08:38.241273+00:00Calibraがロボティクス学習データセットの問題を検出 https://korshunov.ai/ko/article/18023-calibra-robos-hagseub-deiteosesyi-munje-gamji/korshunov.aiko2026-08-12T16:08:38.241273+00:00Calibra, 로봇 학습 데이터셋의 문제 감지 https://korshunov.ai/fr/article/18023-calibra-detecte-des-problemes-dans-les-ensembles-de-donnees-d-apprentissage/korshunov.aifr2026-08-12T16:08:38.241273+00:00Calibra détecte des problèmes dans les ensembles de données d'apprentissage robotique https://korshunov.ai/id/article/18023-calibra-mendeteksi-masalah-dalam-dataset-pembelajaran-robot/korshunov.aiid2026-08-12T16:08:38.241273+00:00Calibra mendeteksi masalah dalam dataset pembelajaran robot https://korshunov.ai/ar/article/18023-tktshf-calibra-mshklt-fy-mjmw-t-bynt-t-lm-lrwbwtt/korshunov.aiar2026-08-12T16:08:38.241273+00:00تكتشف Calibra مشكلات في مجموعات بيانات تعلم الروبوتات https://korshunov.ai/en/article/18023-calibra-detects-issues-in-robot-learning-datasets/korshunov.aien2026-08-12T16:08:38.241273+00:00Calibra detects issues in robot-learning datasets https://korshunov.ai/ru/article/18023-calibra-obnaruzhivaet-problemy-v-naborakh-dannykh-dlia-obucheniia-robotov/korshunov.airu2026-08-12T16:08:38.241273+00:00Calibra обнаруживает проблемы в наборах данных для обучения роботов https://korshunov.ai/es/article/18023-calibra-detecta-problemas-en-conjuntos-de-datos-de-aprendizaje-robotico/korshunov.aies2026-08-12T16:08:38.241273+00:00Calibra detecta problemas en conjuntos de datos de aprendizaje robótico https://korshunov.ai/pt-br/article/18023-calibra-detecta-problemas-em-conjuntos-de-dados-de-aprendizado-robotico/korshunov.aipt2026-08-12T16:08:38.241273+00:00Calibra detecta problemas em conjuntos de dados de aprendizado robótico https://korshunov.ai/hi/article/18023-calibra-robott-lrning-ddettaasett-men-smsyaaon-kaa-ptaa-lgaataa-hai/korshunov.aihi2026-08-12T16:08:38.241273+00:00Calibra रोबोट-लर्निंग डेटासेट में समस्याओं का पता लगाता है https://korshunov.ai/en/article/18015-liquid-ai-releases-lfm2-5-vl-3b-for-faster-edge-vision-capabilities/korshunov.aien2026-08-12T16:00:02.990047+00:00Liquid AI releases LFM2.5-VL-3B for faster edge vision capabilities https://korshunov.ai/ru/article/18015-liquid-ai-vypustila-lfm2-5-vl-3b-dlia-bolee-bystroi-raboty-s-vizualnymi-na/korshunov.airu2026-08-12T16:00:02.990047+00:00Liquid AI выпустила LFM2.5-VL-3B для более быстрой работы с визуальными задачами на периферийных устройствах https://korshunov.ai/es/article/18015-liquid-ai-lanza-lfm2-5-vl-3b-para-capacidades-de-vision-en-el-borde-mas-rapidas/korshunov.aies2026-08-12T16:00:02.990047+00:00Liquid AI lanza LFM2.5-VL-3B para capacidades de visión en el borde más rápidas https://korshunov.ai/zh-hans/article/18015-liquid-ai-fa-bu-lfm2-5-vl-3b-ti-sheng-bian-yuan-shi-jue-neng-li/korshunov.aizh-cn2026-08-12T16:00:02.990047+00:00Liquid AI 发布 LFM2.5-VL-3B，提升边缘视觉能力 https://korshunov.ai/pt-br/article/18015-liquid-ai-lanca-lfm2-5-vl-3b-para-capacidades-de-visao-em-borda-mais-rapidas/korshunov.aipt2026-08-12T16:00:02.990047+00:00Liquid AI lança LFM2.5-VL-3B para capacidades de visão em borda mais rápidas https://korshunov.ai/hi/article/18015-likvidd-eaaii-ne-tej-ej-vijn-kssmtaaon-ke-lie-lfm2-5-vl-3b-jaarii-kiyaa/korshunov.aihi2026-08-12T16:00:02.990047+00:00लिक्विड एआई ने तेज एज विज़न क्षमताओं के लिए LFM2.5-VL-3B जारी किया https://korshunov.ai/ja/article/18015-liquid-ai-kagao-su-naetsushihishiyonji-neng-xiang-keni-lfm2-5-vl-3b-woriri-su/korshunov.aija2026-08-12T16:00:02.990047+00:00Liquid AI が高速なエッジビジョン機能向けに LFM2.5-VL-3B をリリース https://korshunov.ai/ko/article/18015-liquid-aiga-deo-bbareun-esji-bijeon-seongneungeul-wihan-lfm2-5-vl-3b-culsi/korshunov.aiko2026-08-12T16:00:02.990047+00:00Liquid AI가 더 빠른 엣지 비전 성능을 위한 LFM2.5-VL-3B 출시 https://korshunov.ai/fr/article/18015-liquid-ai-publie-lfm2-5-vl-3b-pour-des-capacites-de-vision-sur-peripherique-plus/korshunov.aifr2026-08-12T16:00:02.990047+00:00Liquid AI publie LFM2.5-VL-3B pour des capacités de vision sur périphérique plus rapides https://korshunov.ai/id/article/18015-liquid-ai-merilis-lfm2-5-vl-3b-untuk-kemampuan-visi-tepi-yang-lebih-cepat/korshunov.aiid2026-08-12T16:00:02.990047+00:00Liquid AI merilis LFM2.5-VL-3B untuk kemampuan visi tepi yang lebih cepat https://korshunov.ai/ar/article/18015-liquid-ai-ttlq-lfm2-5-vl-3b-lqdrt-rwy-l-lhf-sr/korshunov.aiar2026-08-12T16:00:02.990047+00:00Liquid AI تطلق LFM2.5-VL-3B لقدرات رؤية على الحافة أسرع https://korshunov.ai/zh-hans/article/18014-google-deepmind-fa-bu-mian-xiang-gboard-he-shi-shi-zhuan-lu-de-sl2t-shou-yu-wen/korshunov.aizh-cn2026-08-12T15:35:52.305605+00:00Google DeepMind 发布面向 Gboard 和实时转录的 SL2T 手语转文本模型 https://korshunov.ai/en/article/18014-google-deepmind-releases-sl2t-sign-language-to-text-model-for-gboard-and-live/korshunov.aien2026-08-12T15:35:52.305605+00:00Google DeepMind releases SL2T sign-language-to-text model for Gboard and Live Transcribe https://korshunov.ai/ru/article/18014-google-deepmind-vypustila-model-perevoda-iazyka-zhestov-v-tekst-sl2t-dlia-gboard/korshunov.airu2026-08-12T15:35:52.305605+00:00Google DeepMind выпустила модель перевода языка жестов в текст SL2T для Gboard и Live Transcribe https://korshunov.ai/es/article/18014-google-deepmind-lanza-el-modelo-sl2t-de-lengua-de-senas-a-texto-para-gboard-y/korshunov.aies2026-08-12T15:35:52.305605+00:00Google DeepMind lanza el modelo SL2T de lengua de señas a texto para Gboard y Live Transcribe https://korshunov.ai/pt-br/article/18014-google-deepmind-lanca-modelo-sl2t-de-traducao-de-lingua-de-sinais-para-texto-no/korshunov.aipt2026-08-12T15:35:52.305605+00:00Google DeepMind lança modelo SL2T de tradução de língua de sinais para texto no Gboard e Live Transcribe https://korshunov.ai/hi/article/18014-google-deepmind-ne-gboard-aur-live-transcribe-ke-lie-sl2t-saain-laingvej-ttuu/korshunov.aihi2026-08-12T15:35:52.305605+00:00Google DeepMind ने Gboard और Live Transcribe के लिए SL2T साइन-लैंग्वेज-टू-टेक्स्ट मॉडल जारी किया https://korshunov.ai/ja/article/18014-google-deepmind-gboardoyohilive-transcribexiang-kenisl2tzhi-wen-zi-huan-kai/korshunov.aija2026-08-12T15:35:52.305605+00:00Google DeepMind、GboardおよびLive Transcribe向けにSL2T指文字からテキストへの変換モデルを公開 https://korshunov.ai/ko/article/18014-google-deepmind-gboard-mic-live-transcribeyong-sl2t-suhwa-tegseuteu-model-culsi/korshunov.aiko2026-08-12T15:35:52.305605+00:00Google DeepMind, Gboard 및 Live Transcribe용 SL2T 수화-텍스트 모델 출시 https://korshunov.ai/fr/article/18014-google-deepmind-publie-le-modele-sl2t-de-traduction-de-langue-des-signes-en-pour/korshunov.aifr2026-08-12T15:35:52.305605+00:00Google DeepMind publie le modèle SL2T de traduction de langue des signes en texte pour Gboard et Live Transcribe https://korshunov.ai/id/article/18014-google-deepmind-merilis-model-sl2t-penerjemah-bahasa-isyarat-ke-teks-untuk-dan/korshunov.aiid2026-08-12T15:35:52.305605+00:00Google DeepMind merilis model SL2T penerjemah bahasa isyarat ke teks untuk Gboard dan Live Transcribe https://korshunov.ai/ar/article/18014-jwjl-dyb-mynd-ttlq-nmwdhj-sl2t-lthwyl-lg-lshr-l-ns-l-gboard-wlive-transcribe/korshunov.aiar2026-08-12T15:35:52.305605+00:00جوجل ديب مايند تطلق نموذج SL2T لتحويل لغة الإشارة إلى نص لـ Gboard وLive Transcribe https://korshunov.ai/en/article/18011-lfm2-5-vl-3b-improves-screen-understanding-and-function-calling-for-edge/korshunov.aien2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B improves screen understanding and function calling for edge deployment https://korshunov.ai/ru/article/18011-lfm2-5-vl-3b-uluchshaet-ponimanie-ekrana-i-vyzov-funktsii-dlia-razvertyvaniia-na/korshunov.airu2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B улучшает понимание экрана и вызов функций для развёртывания на периферии https://korshunov.ai/es/article/18011-lfm2-5-vl-3b-mejora-la-comprension-de-pantallas-y-la-invocacion-de-funciones-en/korshunov.aies2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B mejora la comprensión de pantallas y la invocación de funciones para despliegue en el borde https://korshunov.ai/zh-hans/article/18011-lfm2-5-vl-3bti-sheng-ping-mu-li-jie-yu-han-shu-diao-yong-neng-li-zhu-li-bian-bu/korshunov.aizh-cn2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B提升屏幕理解与函数调用能力，助力边缘部署 https://korshunov.ai/pt-br/article/18011-lfm2-5-vl-3b-melhora-a-compreensao-de-tela-e-a-chamada-de-funcoes-para-em-de/korshunov.aipt2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B melhora a compreensão de tela e a chamada de funções para implantação em dispositivos de borda https://korshunov.ai/hi/article/18011-lfm2-5-vl-3b-ej-ddiplonymentt-ke-lie-skriin-smjh-aur-phnkshn-konling-men-sudhaar/korshunov.aihi2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B एज डिप्लॉयमेंट के लिए स्क्रीन समझ और फंक्शन कॉलिंग में सुधार करता है https://korshunov.ai/ja/article/18011-lfm2-5-vl-3bkaetsushizhan-kai-xiang-kenohua-mian-li-jie-toguan-shu-hu-hichu-shan/korshunov.aija2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3Bがエッジ展開向けの画面理解と関数呼び出しを改善 https://korshunov.ai/ko/article/18011-lfm2-5-vl-3bga-esji-baeporeul-wihan-hwamyeon-ihae-mic-hamsu-hoculeul-gaeseon/korshunov.aiko2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B가 엣지 배포를 위한 화면 이해 및 함수 호출을 개선 https://korshunov.ai/fr/article/18011-lfm2-5-vl-3b-ameliore-la-comprehension-d-ecran-et-l-appel-de-fonctions-pour-le/korshunov.aifr2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B améliore la compréhension d'écran et l'appel de fonctions pour le déploiement en périphérie https://korshunov.ai/id/article/18011-lfm2-5-vl-3b-meningkatkan-pemahaman-layar-dan-pemanggilan-fungsi-untuk-di-tepi/korshunov.aiid2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B meningkatkan pemahaman layar dan pemanggilan fungsi untuk penyebaran di perangkat tepi https://korshunov.ai/ar/article/18011-lfm2-5-vl-3b-yhswn-fhm-lshsh-wstd-lwzyf-llnshr-l-lhf/korshunov.aiar2026-08-12T14:27:32.037482+00:00LFM2.5-VL-3B يحسّن فهم الشاشة واستدعاء الوظائف للنشر على الحافة https://korshunov.ai/en/article/18009-openai-reports-widening-gap-as-enterprise-ai-shifts-from-assistance-to-execution/korshunov.aien2026-08-12T14:00:41.960095+00:00OpenAI reports widening gap as enterprise AI shifts from assistance to execution https://korshunov.ai/ru/article/18009-openai-soobshchaet-o-rasshirenii-razryva-po-mere-perekhoda-korporativnogo-ii-ot/korshunov.airu2026-08-12T14:00:41.960095+00:00OpenAI сообщает о расширении разрыва по мере перехода корпоративного ИИ от помощи к выполнению задач https://korshunov.ai/es/article/18009-openai-informa-de-una-brecha-creciente-a-medida-que-la-ia-empresarial-pasa-de-a/korshunov.aies2026-08-12T14:00:41.960095+00:00OpenAI informa de una brecha creciente a medida que la IA empresarial pasa de asistencia a ejecución https://korshunov.ai/zh-hans/article/18009-openaibao-gao-cheng-qi-ye-aicong-fu-zhu-zhuan-xiang-zhi-xing-chai-ju-kuo-da/korshunov.aizh-cn2026-08-12T14:00:41.960095+00:00OpenAI报告称企业AI从辅助转向执行，差距扩大 https://korshunov.ai/pt-br/article/18009-openai-relata-ampliacao-da-lacuna-a-medida-que-a-ia-empresarial-migra-de-para/korshunov.aipt2026-08-12T14:00:41.960095+00:00OpenAI relata ampliação da lacuna à medida que a IA empresarial migra de assistência para execução https://korshunov.ai/hi/article/18009-udym-eaaii-shaaytaa-se-nisspaadn-kii-or-bddh-rhaa-hai-opneaaii-antr-ke-bddhne/korshunov.aihi2026-08-12T14:00:41.960095+00:00उद्यम एआई सहायता से निष्पादन की ओर बढ़ रहा है, ओपनएआई अंतर के बढ़ने की रिपोर्ट करता है https://korshunov.ai/ja/article/18009-openai-enta-huraisuaikazhi-yuan-karashi-xing-heyi-xing-suruzhong-tege-chai-kuo/korshunov.aija2026-08-12T14:00:41.960095+00:00OpenAI、エンタープライズAIが支援から実行へ移行する中で格差拡大を報告 https://korshunov.ai/ko/article/18009-openai-gieobyong-aiga-bojoeseo-silhaengeuro-jeonhwandoemyeonseo-gyeogca-hwagdae/korshunov.aiko2026-08-12T14:00:41.960095+00:00OpenAI, 기업용 AI가 보조에서 실행으로 전환되면서 격차 확대 보고 https://korshunov.ai/fr/article/18009-openai-signale-un-ecart-croissant-alors-que-l-ia-d-entreprise-passe-de-l-a-l/korshunov.aifr2026-08-12T14:00:41.960095+00:00OpenAI signale un écart croissant alors que l'IA d'entreprise passe de l'assistance à l'exécution https://korshunov.ai/id/article/18009-openai-melaporkan-kesenjangan-yang-melebar-seiring-pergeseran-ai-perusahaan-dari/korshunov.aiid2026-08-12T14:00:41.960095+00:00OpenAI melaporkan kesenjangan yang melebar seiring pergeseran AI perusahaan dari bantuan ke eksekusi https://korshunov.ai/ar/article/18009-openai-tublg-n-ts-lfjw-m-thwl-ldhk-lstn-y-lmwssy-mn-lms-d-l-ltnfydh/korshunov.aiar2026-08-12T14:00:41.960095+00:00OpenAI تُبلغ عن اتساع الفجوة مع تحول الذكاء الاصطناعي المؤسسي من المساعدة إلى التنفيذ https://korshunov.ai/zh-hans/article/17996-llmzai-chao-can-shu-you-hua-zhong-pi-pei-huo-you-yu-bei-xie-si-you-hua/korshunov.aizh-cn2026-08-12T11:13:55.389105+00:00LLM在超参数优化中匹配或优于贝叶斯优化 https://korshunov.ai/pt-br/article/17996-llms-igualam-ou-superam-a-otimizacao-bayesiana-para-otimizacao-de/korshunov.aipt2026-08-12T11:13:55.389105+00:00LLMs igualam ou superam a otimização bayesiana para otimização de hiperparâmetros https://korshunov.ai/hi/article/17996-llm-haaiprpairaamiittr-onpttimaaijeshn-men-beysiyn-onpttimaaijeshn-ke-braabr-yaa/korshunov.aihi2026-08-12T11:13:55.389105+00:00LLM हाइपरपैरामीटर ऑप्टिमाइज़ेशन में बेयसियन ऑप्टिमाइज़ेशन के बराबर या बेहतर हैं https://korshunov.ai/ja/article/17996-llmhahaiha-harame-tazui-shi-hua-nioiteheisuzui-shi-hua-totong-deng-kasorewochao/korshunov.aija2026-08-12T11:13:55.389105+00:00LLMはハイパーパラメータ最適化においてベイズ最適化と同等かそれを超える https://korshunov.ai/ko/article/17996-llmi-haipeoparamiteo-coejeoghwaeseo-beijian-coejeoghwawa-dongdeunghageona-ireul/korshunov.aiko2026-08-12T11:13:55.389105+00:00LLM이 하이퍼파라미터 최적화에서 베이지안 최적화와 동등하거나 이를 능가함 https://korshunov.ai/en/article/17996-llms-match-or-beat-bayesian-optimization-for-hyperparameter-optimization/korshunov.aien2026-08-12T11:13:55.389105+00:00LLMs match or beat Bayesian optimization for hyperparameter optimization https://korshunov.ai/ru/article/17996-llm-sootvetstvuiut-ili-prevoskhodiat-baiesovskuiu-optimizatsiiu-v-optimizatsii/korshunov.airu2026-08-12T11:13:55.389105+00:00LLM соответствуют или превосходят байесовскую оптимизацию в гиперпараметрической оптимизации https://korshunov.ai/es/article/17996-los-llm-igualan-o-superan-la-optimizacion-bayesiana-para-la-optimizacion-de/korshunov.aies2026-08-12T11:13:55.389105+00:00Los LLM igualan o superan la optimización bayesiana para la optimización de hiperparámetros https://korshunov.ai/fr/article/17996-les-llm-egalent-ou-surpassent-l-optimisation-bayesienne-pour-l-optimisation-des/korshunov.aifr2026-08-12T11:13:55.389105+00:00Les LLM égalent ou surpassent l'optimisation bayésienne pour l'optimisation des hyperparamètres https://korshunov.ai/id/article/17996-llm-menyamai-atau-mengalahkan-optimisasi-bayesian-untuk-optimisasi/korshunov.aiid2026-08-12T11:13:55.389105+00:00LLM menyamai atau mengalahkan optimisasi Bayesian untuk optimisasi hiperparameter https://korshunov.ai/ar/article/17996-tswy-nmdhj-llgt-lkbyr-w-ttfwq-l-lthsyn-lbyzy-fy-thsyn-lm-mlt-lfyq/korshunov.aiar2026-08-12T11:13:55.389105+00:00تساوي نماذج اللغات الكبيرة أو تتفوق على التحسين البايزي في تحسين المعاملات الفائقة https://korshunov.ai/zh-hans/article/17995-sourceyin-ru-jin-si-zhan-kai-wei-fen-yong-yu-xun-lian-shu-ju-gui-yin/korshunov.aizh-cn2026-08-12T11:06:14.407992+00:00Source引入近似展开微分用于训练数据归因 https://korshunov.ai/en/article/17995-source-introduces-approximate-unrolled-differentiation-for-training-data/korshunov.aien2026-08-12T11:06:14.407992+00:00Source introduces approximate unrolled differentiation for training data attribution https://korshunov.ai/ru/article/17995-istochnik-predstavliaet-priblizhennoe-razvernutoe-differentsirovanie-dlia/korshunov.airu2026-08-12T11:06:14.407992+00:00Источник представляет приближённое развёрнутое дифференцирование для атрибуции обучающих данных https://korshunov.ai/es/article/17995-source-introduce-diferenciacion-desenrollada-aproximada-para-la-atribucion-de-de/korshunov.aies2026-08-12T11:06:14.407992+00:00Source introduce diferenciación desenrollada aproximada para la atribución de datos de entrenamiento https://korshunov.ai/pt-br/article/17995-fonte-introduz-diferenciacao-desenrolada-aproximada-para-atribuicao-de-dados-de/korshunov.aipt2026-08-12T11:06:14.407992+00:00Fonte introduz diferenciação desenrolada aproximada para atribuição de dados de treinamento https://korshunov.ai/hi/article/17995-srot-prshikssnn-ddettaa-ettribyuushn-ke-lie-anumaanit-anroldd-ddiphrenshieshn/korshunov.aihi2026-08-12T11:06:14.407992+00:00स्रोत प्रशिक्षण डेटा एट्रिब्यूशन के लिए अनुमानित अनरोल्ड डिफरेंशिएशन पेश करता है https://korshunov.ai/ja/article/17995-sourcekatore-ninkute-tashu-xing-fu-kenotamenojin-si-anro-ruwei-fen-wodao-ru/korshunov.aija2026-08-12T11:06:14.407992+00:00Sourceがトレーニングデータ属性付けのための近似アンロール微分を導入 https://korshunov.ai/ko/article/17995-sourcega-hunryeon-deiteo-sogseong-cujeongeul-wihan-geunsa-eonrolring-mibun-doib/korshunov.aiko2026-08-12T11:06:14.407992+00:00Source가 훈련 데이터 속성 추정을 위한 근사 언롤링 미분 도입 https://korshunov.ai/fr/article/17995-source-introduit-la-differenciation-deroulee-approximative-pour-l-attribution-d/korshunov.aifr2026-08-12T11:06:14.407992+00:00Source introduit la différenciation déroulée approximative pour l'attribution des données d'entraînement https://korshunov.ai/id/article/17995-source-memperkenalkan-diferensiasi-unrolled-aproksimasi-untuk-atribusi-data/korshunov.aiid2026-08-12T11:06:14.407992+00:00Source memperkenalkan diferensiasi unrolled aproksimasi untuk atribusi data pelatihan https://korshunov.ai/ar/article/17995-msdr-yqdm-shtqqan-mqrban-gyr-mltwy-lnsb-bynt-ltdryb/korshunov.aiar2026-08-12T11:06:14.407992+00:00مصدر يقدم اشتقاقًا مقاربًا غير ملتوي لنسبة بيانات التدريب https://korshunov.ai/zh-hans/article/17994-nvidiafa-bu-omnidreamsshi-shi-sheng-cheng-shi-shi-jie-mo-xing-yong-yu-zi-dong/korshunov.aizh-cn2026-08-12T10:59:43.423527+00:00NVIDIA发布OmniDreams实时生成式世界模型，用于自动驾驶仿真 https://korshunov.ai/en/article/17994-nvidia-releases-omnidreams-real-time-generative-world-model-for-autonomous/korshunov.aien2026-08-12T10:59:43.423527+00:00NVIDIA releases OmniDreams real-time generative world model for autonomous vehicle simulation https://korshunov.ai/ru/article/17994-nvidia-vypustila-model-generatsii-mira-v-realnom-vremeni-omnidreams-dlia-sredstv/korshunov.airu2026-08-12T10:59:43.423527+00:00NVIDIA выпустила модель генерации мира в реальном времени OmniDreams для симуляции автономных транспортных средств https://korshunov.ai/es/article/17994-nvidia-lanza-omnidreams-un-modelo-generativo-de-mundo-en-tiempo-real-para-la-de/korshunov.aies2026-08-12T10:59:43.423527+00:00NVIDIA lanza OmniDreams, un modelo generativo de mundo en tiempo real para la simulación de vehículos autónomos https://korshunov.ai/pt-br/article/17994-nvidia-lanca-omnidreams-modelo-de-mundo-generativo-em-tempo-real-para-simulacao/korshunov.aipt2026-08-12T10:59:43.423527+00:00NVIDIA lança OmniDreams, modelo de mundo generativo em tempo real para simulação de veículos autônomos https://korshunov.ai/hi/article/17994-svcaalit-vaahn-simuleshn-ke-lie-nvidia-ne-omnidreams-riyl-ttaaim-jnrettiv-vrldd/korshunov.aihi2026-08-12T10:59:43.423527+00:00स्वचालित वाहन सिमुलेशन के लिए NVIDIA ने OmniDreams रियल-टाइम जनरेटिव वर्ल्ड मॉडल जारी किया https://korshunov.ai/ja/article/17994-nvidiakazi-dong-yun-zhuan-shimiyure-shiyonxiang-keriarutaimusheng-cheng-xing-shi/korshunov.aija2026-08-12T10:59:43.423527+00:00NVIDIAが自動運転シミュレーション向けリアルタイム生成型世界モデル「OmniDreams」をリリース https://korshunov.ai/ko/article/17994-nvidia-jayuljuhaeng-simyulreisyeoneul-wihan-silsigan-saengseonghyeong-weoldeu/korshunov.aiko2026-08-12T10:59:43.423527+00:00NVIDIA, 자율주행 시뮬레이션을 위한 실시간 생성형 월드 모델 OmniDreams 출시 https://korshunov.ai/fr/article/17994-nvidia-publie-omnidreams-un-modele-de-monde-generatif-en-temps-reel-pour-la-de/korshunov.aifr2026-08-12T10:59:43.423527+00:00NVIDIA publie OmniDreams, un modèle de monde génératif en temps réel pour la simulation de véhicules autonomes https://korshunov.ai/id/article/17994-nvidia-merilis-model-dunia-generatif-waktu-nyata-omnidreams-untuk-simulasi/korshunov.aiid2026-08-12T10:59:43.423527+00:00NVIDIA merilis model dunia generatif waktu nyata OmniDreams untuk simulasi kendaraan otonom https://korshunov.ai/ar/article/17994-nfydy-ttlq-nmwdhj-l-lm-ltwlydy-fy-lwqt-lf-ly-omnidreams-lmhk-lmrkbt-dhty-lqyd/korshunov.aiar2026-08-12T10:59:43.423527+00:00إنفيديا تطلق نموذج العالم التوليدي في الوقت الفعلي OmniDreams لمحاكاة المركبات ذاتية القيادة https://korshunov.ai/ja/article/17993-jacnethagou-zao-hua-yakohianwozhi-jie-moteruhua-surukototeguan-shu-woxue-xi-suru/korshunov.aija2026-08-12T10:40:21.833379+00:00JacNetは構造化ヤコビアンを直接モデル化することで関数を学習する https://korshunov.ai/ko/article/17993-jacneteun-gujohwadoen-yakobianeul-jigjeob-modelringhayeo-hamsureul/korshunov.aiko2026-08-12T10:40:21.833379+00:00JacNet은 구조화된 야코비안을 직접 모델링하여 함수를 학습합니다 https://korshunov.ai/fr/article/17993-jacnet-apprend-des-fonctions-en-modelisant-directement-les-jacobiens-structures/korshunov.aifr2026-08-12T10:40:21.833379+00:00JacNet apprend des fonctions en modélisant directement les jacobiens structurés https://korshunov.ai/id/article/17993-jacnet-mempelajari-fungsi-dengan-memodelkan-langsung-jacobian-terstruktur/korshunov.aiid2026-08-12T10:40:21.833379+00:00JacNet mempelajari fungsi dengan memodelkan langsung Jacobian terstruktur https://korshunov.ai/ar/article/17993-tt-lm-jacnet-ldwl-n-tryq-nmdhj-jkwbyn-lmhykl-mbshr/korshunov.aiar2026-08-12T10:40:21.833379+00:00تتعلم JacNet الدوال عن طريق نمذجة جاكوبيان المهيكلة مباشرة https://korshunov.ai/en/article/17993-jacnet-learns-functions-by-directly-modeling-structured-jacobians/korshunov.aien2026-08-12T10:40:21.833379+00:00JacNet learns functions by directly modeling structured Jacobians https://korshunov.ai/ru/article/17993-jacnet-uchit-funktsii-putem-priamogo-modelirovaniia-strukturirovannykh/korshunov.airu2026-08-12T10:40:21.833379+00:00JacNet учит функции путём прямого моделирования структурированных якобианов https://korshunov.ai/es/article/17993-jacnet-aprende-funciones-modelando-directamente-jacobianos-estructurados/korshunov.aies2026-08-12T10:40:21.833379+00:00JacNet aprende funciones modelando directamente jacobianos estructurados https://korshunov.ai/zh-hans/article/17993-jacnettong-guo-zhi-jie-jian-mo-jie-gou-hua-ya-ke-bi-ju-zhen-lai-xue-xi-han-shu/korshunov.aizh-cn2026-08-12T10:40:21.833379+00:00JacNet通过直接建模结构化雅可比矩阵来学习函数 https://korshunov.ai/pt-br/article/17993-jacnet-aprende-funcoes-modelando-diretamente-jacobianos-estruturados/korshunov.aipt2026-08-12T10:40:21.833379+00:00JacNet aprende funções modelando diretamente Jacobianos estruturados https://korshunov.ai/hi/article/17993-jacnet-snrcit-jaikobiyn-ko-siidhe-monddl-krke-phnkshns-siikhtaa-hai/korshunov.aihi2026-08-12T10:40:21.833379+00:00JacNet संरचित जैकोबियन को सीधे मॉडल करके फ़ंक्शंस सीखता है https://korshunov.ai/fr/article/17992-jlorraine-presente-des-outils-d-optimisation-imbriquee-evolutifs-pour-l-profond/korshunov.aifr2026-08-12T10:33:46.843314+00:00jlorraine présente des outils d'optimisation imbriquée évolutifs pour l'apprentissage profond https://korshunov.ai/id/article/17992-jlorraine-menyajikan-alat-optimasi-bersarang-yang-skalabel-untuk-pembelajaran/korshunov.aiid2026-08-12T10:33:46.843314+00:00jlorraine menyajikan alat optimasi bersarang yang skalabel untuk pembelajaran mendalam https://korshunov.ai/ar/article/17992-jlorraine-tqdm-dwt-thsyn-mtdkhl-qbl-lltws-llt-lm-l-myq/korshunov.aiar2026-08-12T10:33:46.843314+00:00jlorraine تقدم أدوات تحسين متداخلة قابلة للتوسع للتعلم العميق https://korshunov.ai/en/article/17992-jlorraine-presents-scalable-nested-optimization-tools-for-deep-learning/korshunov.aien2026-08-12T10:33:46.843314+00:00jlorraine presents scalable nested optimization tools for deep learning https://korshunov.ai/ru/article/17992-jlorraine-predstavliaet-masshtabiruemye-instrumenty-vlozhennoi-optimizatsii-dlia/korshunov.airu2026-08-12T10:33:46.843314+00:00jlorraine представляет масштабируемые инструменты вложенной оптимизации для глубокого обучения https://korshunov.ai/es/article/17992-jlorraine-presenta-herramientas-de-optimizacion-anidada-escalables-para-el/korshunov.aies2026-08-12T10:33:46.843314+00:00jlorraine presenta herramientas de optimización anidada escalables para el aprendizaje profundo https://korshunov.ai/zh-hans/article/17992-jlorraine-ti-chu-yong-yu-shen-du-xue-xi-de-ke-kuo-zhan-qian-tao-you-hua-gong-ju/korshunov.aizh-cn2026-08-12T10:33:46.843314+00:00jlorraine 提出用于深度学习的可扩展嵌套优化工具 https://korshunov.ai/pt-br/article/17992-jlorraine-apresenta-ferramentas-de-otimizacao-aninhada-escalaveis-para-profundo/korshunov.aipt2026-08-12T10:33:46.843314+00:00jlorraine apresenta ferramentas de otimização aninhada escaláveis para aprendizado profundo https://korshunov.ai/hi/article/17992-jlorraine-ghn-shikssnn-ke-lie-skelebl-nesttedd-onpttimaaijeshn-ttuuls-prstut-hai/korshunov.aihi2026-08-12T10:33:46.843314+00:00jlorraine गहन शिक्षण के लिए स्केलेबल नेस्टेड ऑप्टिमाइजेशन टूल्स प्रस्तुत करता है https://korshunov.ai/ja/article/17992-jlorraine-katei-hura-ninkuyong-nosuke-rahurunanesutozui-shi-hua-tsu-ruwofa-biao/korshunov.aija2026-08-12T10:33:46.843314+00:00jlorraine がディープラーニング用のスケーラブルなネスト最適化ツールを発表 https://korshunov.ai/ko/article/17992-jlorraine-dibreoningeul-wihan-hwagjang-ganeunghan-jungceob-coejeoghwa-dogu/korshunov.aiko2026-08-12T10:33:46.843314+00:00jlorraine, 딥러닝을 위한 확장 가능한 중첩 최적화 도구 발표 https://korshunov.ai/en/article/17988-nvidia-releases-nemotron-3-5-lightning-model-and-nemo-switchyard-router/korshunov.aien2026-08-12T10:01:02.037316+00:00NVIDIA releases Nemotron 3.5 Lightning model and NeMo Switchyard router https://korshunov.ai/ru/article/17988-nvidia-vypustila-model-nemotron-3-5-lightning-i-marshrutizator-nemo-switchyard/korshunov.airu2026-08-12T10:01:02.037316+00:00NVIDIA выпустила модель Nemotron 3.5 Lightning и маршрутизатор NeMo Switchyard https://korshunov.ai/es/article/17988-nvidia-lanza-el-modelo-nemotron-3-5-lightning-y-el-enrutador-nemo-switchyard/korshunov.aies2026-08-12T10:01:02.037316+00:00NVIDIA lanza el modelo Nemotron 3.5 Lightning y el enrutador NeMo Switchyard https://korshunov.ai/zh-hans/article/17988-nvidiafa-bu-nemotron-3-5-lightningmo-xing-he-nemo-switchyardlu-you-qi/korshunov.aizh-cn2026-08-12T10:01:02.037316+00:00NVIDIA发布Nemotron 3.5 Lightning模型和NeMo Switchyard路由器 https://korshunov.ai/pt-br/article/17988-nvidia-lanca-modelo-nemotron-3-5-lightning-e-roteador-nemo-switchyard/korshunov.aipt2026-08-12T10:01:02.037316+00:00NVIDIA lança modelo Nemotron 3.5 Lightning e roteador NeMo Switchyard https://korshunov.ai/hi/article/17988-nvidia-ne-nemotron-3-5-lightning-monddl-aur-nemo-switchyard-raauttr-jaarii-kie/korshunov.aihi2026-08-12T10:01:02.037316+00:00NVIDIA ने Nemotron 3.5 Lightning मॉडल और NeMo Switchyard राउटर जारी किए https://korshunov.ai/ja/article/17988-nvidia-nemotron-3-5-lightningmoterutonemo-switchyardru-ta-woriri-su/korshunov.aija2026-08-12T10:01:02.037316+00:00NVIDIA、Nemotron 3.5 LightningモデルとNeMo Switchyardルーターをリリース https://korshunov.ai/ko/article/17988-nvidiaga-nemotron-3-5-lightning-modelgwa-nemo-switchyard-rauteoreul-culsihada/korshunov.aiko2026-08-12T10:01:02.037316+00:00NVIDIA가 Nemotron 3.5 Lightning 모델과 NeMo Switchyard 라우터를 출시하다 https://korshunov.ai/fr/article/17988-nvidia-publie-le-modele-nemotron-3-5-lightning-et-le-routeur-nemo-switchyard/korshunov.aifr2026-08-12T10:01:02.037316+00:00NVIDIA publie le modèle Nemotron 3.5 Lightning et le routeur NeMo Switchyard https://korshunov.ai/id/article/17988-nvidia-merilis-model-nemotron-3-5-lightning-dan-router-nemo-switchyard/korshunov.aiid2026-08-12T10:01:02.037316+00:00NVIDIA merilis model Nemotron 3.5 Lightning dan router NeMo Switchyard https://korshunov.ai/ar/article/17988-nfydy-ttlq-nmwdhj-nemotron-3-5-lightning-wmwjwh-nemo-switchyard/korshunov.aiar2026-08-12T10:01:02.037316+00:00إنفيديا تطلق نموذج Nemotron 3.5 Lightning وموجّه NeMo Switchyard https://korshunov.ai/en/article/17982-researchers-decode-encrypted-reasoning-traces-from-claude-gpt-and-gemini-apis/korshunov.aien2026-08-12T08:12:56.226865+00:00Researchers decode encrypted reasoning traces from Claude, GPT, and Gemini APIs https://korshunov.ai/ru/article/17982-issledovateli-rasshifrovali-tsepochki-rassuzhdenii-iz-api-claude-gpt-i-gemini/korshunov.airu2026-08-12T08:12:56.226865+00:00Исследователи расшифровали цепочки рассуждений из API Claude, GPT и Gemini https://korshunov.ai/es/article/17982-investigadores-descifran-trazas-de-razonamiento-cifradas-de-las-apis-de-claude-y/korshunov.aies2026-08-12T08:12:56.226865+00:00Investigadores descifran trazas de razonamiento cifradas de las APIs de Claude, GPT y Gemini https://korshunov.ai/zh-hans/article/17982-yan-jiu-ren-yuan-jie-ma-lai-zi-claude-gpt-he-gemini-api-de-jia-mi-tui-li-gui-ji/korshunov.aizh-cn2026-08-12T08:12:56.226865+00:00研究人员解码来自 Claude、GPT 和 Gemini API 的加密推理轨迹 https://korshunov.ai/pt-br/article/17982-pesquisadores-decodificam-rastros-de-raciocinio-criptografados-das-apis-claude-e/korshunov.aipt2026-08-12T08:12:56.226865+00:00Pesquisadores decodificam rastros de raciocínio criptografados das APIs Claude, GPT e Gemini https://korshunov.ai/hi/article/17982-shodhkrtaaon-ne-klondd-jiipiittii-aur-jeminii-epiiaaii-se-enkripttedd-trk-ttres/korshunov.aihi2026-08-12T08:12:56.226865+00:00शोधकर्ताओं ने क्लॉड, जीपीटी और जेमिनी एपीआई से एन्क्रिप्टेड तर्क ट्रेस को डीकोड किया https://korshunov.ai/ja/article/17982-yan-jiu-zhe-rakaclaude-gpt-gemininoapikaraan-hao-hua-saretatui-lun-tore-suwojie/korshunov.aija2026-08-12T08:12:56.226865+00:00研究者らがClaude、GPT、GeminiのAPIから暗号化された推論トレースを解読 https://korshunov.ai/ko/article/17982-yeongujadeuli-claude-gpt-gemini-apiyi-amhohwadoen-curon-teureiseureul-haedoghada/korshunov.aiko2026-08-12T08:12:56.226865+00:00연구자들이 Claude, GPT, Gemini API의 암호화된 추론 트레이스를 해독하다 https://korshunov.ai/fr/article/17982-des-chercheurs-decryptent-les-traces-de-raisonnement-chiffrees-des-api-claude-et/korshunov.aifr2026-08-12T08:12:56.226865+00:00Des chercheurs décryptent les traces de raisonnement chiffrées des API Claude, GPT et Gemini https://korshunov.ai/id/article/17982-peneliti-menguraikan-jejak-penalaran-terenkripsi-dari-api-claude-gpt-dan-gemini/korshunov.aiid2026-08-12T08:12:56.226865+00:00Peneliti menguraikan jejak penalaran terenkripsi dari API Claude, GPT, dan Gemini https://korshunov.ai/ar/article/17982-bhthwn-yfkwwn-tshfyr-msrt-lstdll-mn-wjht-brmj-ttbyqt-claude-wgpt-wgemini/korshunov.aiar2026-08-12T08:12:56.226865+00:00باحثون يفكّون تشفير مسارات الاستدلال من واجهات برمجة تطبيقات Claude وGPT وGemini https://korshunov.ai/pt-br/article/17979-llama-cpp-b10369-adiciona-suporte-ao-pocket-tts-com-geracao-cuda-80-mais-rapida/korshunov.aipt2026-08-12T05:34:13.655825+00:00llama.cpp b10369 adiciona suporte ao pocket-tts com geração CUDA 80% mais rápida https://korshunov.ai/hi/article/17979-llama-cpp-b10369-men-pocket-tts-smrthn-joddaa-gyaa-hai-cuda-jnreshn-80-tej/korshunov.aihi2026-08-12T05:34:13.655825+00:00llama.cpp b10369 में pocket-tts समर्थन जोड़ा गया है, CUDA जनरेशन 80% तेज https://korshunov.ai/ja/article/17979-llama-cpp-b10369kapocket-ttssaho-towozhui-jia-shi-cudasheng-cheng-ka80-gao-su/korshunov.aija2026-08-12T05:34:13.655825+00:00llama.cpp b10369がpocket-ttsサポートを追加し、CUDA生成が80%高速化 https://korshunov.ai/ko/article/17979-llama-cpp-b10369-pocket-tts-jiweon-cuga-mic-cuda-saengseong-80-gasog/korshunov.aiko2026-08-12T05:34:13.655825+00:00llama.cpp b10369, pocket-tts 지원 추가 및 CUDA 생성 80% 가속 https://korshunov.ai/fr/article/17979-llama-cpp-b10369-ajoute-le-support-de-pocket-tts-avec-une-generation-cuda-80/korshunov.aifr2026-08-12T05:34:13.655825+00:00llama.cpp b10369 ajoute le support de pocket-tts avec une génération CUDA 80% plus rapide https://korshunov.ai/id/article/17979-llama-cpp-b10369-menambahkan-dukungan-pocket-tts-dengan-generasi-cuda-80-lebih/korshunov.aiid2026-08-12T05:34:13.655825+00:00llama.cpp b10369 menambahkan dukungan pocket-tts dengan generasi CUDA 80% lebih cepat https://korshunov.ai/ar/article/17979-df-llama-cpp-b10369-d-m-pocket-tts-m-tsry-twlyd-cuda-bnsb-80/korshunov.aiar2026-08-12T05:34:13.655825+00:00إضافة llama.cpp b10369 دعم pocket-tts مع تسريع توليد CUDA بنسبة 80% https://korshunov.ai/zh-hans/article/17979-llama-cpp-b10369-zeng-jia-pocket-tts-zhi-chi-cuda-sheng-cheng-su-du-ti-sheng-80/korshunov.aizh-cn2026-08-12T05:34:13.655825+00:00llama.cpp b10369 增加 pocket-tts 支持，CUDA 生成速度提升 80% https://korshunov.ai/en/article/17979-llama-cpp-b10369-adds-pocket-tts-support-with-80-faster-cuda-generation/korshunov.aien2026-08-12T05:34:13.655825+00:00llama.cpp b10369 adds pocket-tts support with 80% faster CUDA generation https://korshunov.ai/ru/article/17979-llama-cpp-b10369-dobavliaet-podderzhku-pocket-tts-s-uskoreniem-cuda-generatsii/korshunov.airu2026-08-12T05:34:13.655825+00:00llama.cpp b10369 добавляет поддержку pocket-tts с ускорением CUDA-генерации на 80% https://korshunov.ai/es/article/17979-llama-cpp-b10369-anade-soporte-para-pocket-tts-con-generacion-cuda-un-80-mas/korshunov.aies2026-08-12T05:34:13.655825+00:00llama.cpp b10369 añade soporte para pocket-tts con generación CUDA un 80% más rápida https://korshunov.ai/en/article/17912-milmmt-46-v1-0-improves-multilingual-translation-via-reference-free-post/korshunov.aien2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 improves multilingual translation via reference-free post-training https://korshunov.ai/ru/article/17912-milmmt-46-v1-0-uluchshaet-mnogoiazychnyi-perevod-s-pomoshchiu-postobucheniia-bez/korshunov.airu2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 улучшает многоязычный перевод с помощью постобучения без эталона https://korshunov.ai/es/article/17912-milmmt-46-v1-0-mejora-la-traduccion-multilingue-mediante-postentrenamiento-sin/korshunov.aies2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 mejora la traducción multilingüe mediante postentrenamiento sin referencia https://korshunov.ai/zh-hans/article/17912-milmmt-46-v1-0-tong-guo-wu-can-kao-hou-xun-lian-ti-sheng-duo-yu-yan-fan-yi-neng/korshunov.aizh-cn2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 通过无参考后训练提升多语言翻译能力 https://korshunov.ai/pt-br/article/17912-milmmt-46-v1-0-melhora-a-traducao-multilingue-por-meio-de-pos-treinamento-sem/korshunov.aipt2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 melhora a tradução multilíngue por meio de pós-treinamento sem referência https://korshunov.ai/hi/article/17912-sndrbh-mukt-postt-ttrening-ke-maadhym-se-milmmt-46-v1-0-bhubhaassii-anuvaad-men/korshunov.aihi2026-08-12T04:47:44.472650+00:00संदर्भ-मुक्त पोस्ट-ट्रेनिंग के माध्यम से MiLMMT-46-v1.0 बहुभाषी अनुवाद में सुधार करता है https://korshunov.ai/ja/article/17912-milmmt-46-v1-0kacan-zhao-nashihou-xun-lian-niyoriduo-yan-yu-fan-yi-wogai-shan/korshunov.aija2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0が参照なし後訓練により多言語翻訳を改善 https://korshunov.ai/ko/article/17912-milmmt-46-v1-0-repeoreonseu-eobsneun-sahu-hagseubeul-tonghan-dagugeo-beonyeog/korshunov.aiko2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0, 레퍼런스 없는 사후 학습을 통한 다국어 번역 성능 향상 https://korshunov.ai/fr/article/17912-milmmt-46-v1-0-ameliore-la-traduction-multilingue-grace-a-un-post-entrainement/korshunov.aifr2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 améliore la traduction multilingue grâce à un post-entraînement sans référence https://korshunov.ai/id/article/17912-milmmt-46-v1-0-meningkatkan-terjemahan-multibahasa-melalui-pasca-pelatihan-bebas/korshunov.aiid2026-08-12T04:47:44.472650+00:00MiLMMT-46-v1.0 meningkatkan terjemahan multibahasa melalui pasca-pelatihan bebas referensi https://korshunov.ai/ar/article/17912-yhswn-milmmt-46-v1-0-ltrjm-mt-dd-llgt-br-m-b-d-ltdryb-bdwn-mrj/korshunov.aiar2026-08-12T04:47:44.472650+00:00يحسّن MiLMMT-46-v1.0 الترجمة متعددة اللغات عبر ما بعد التدريب بدون مرجع https://korshunov.ai/en/article/17971-the-ai-breakroom-lets-agents-build-culture-in-shared-chat-rooms/korshunov.aien2026-08-12T03:29:17.741368+00:00The AI Breakroom lets agents build culture in shared chat rooms https://korshunov.ai/ru/article/17971-ai-breakroom-pozvoliaet-agentam-formirovat-kulturu-v-obshchikh-chatakh/korshunov.airu2026-08-12T03:29:17.741368+00:00AI Breakroom позволяет агентам формировать культуру в общих чатах https://korshunov.ai/es/article/17971-ai-breakroom-permite-a-los-agentes-construir-cultura-en-salas-de-chat/korshunov.aies2026-08-12T03:29:17.741368+00:00AI Breakroom permite a los agentes construir cultura en salas de chat compartidas https://korshunov.ai/zh-hans/article/17971-ai-breakroom-rang-zhi-neng-ti-zai-gong-xiang-liao-tian-shi-zhong-gou-jian-wen/korshunov.aizh-cn2026-08-12T03:29:17.741368+00:00AI Breakroom 让智能体在共享聊天室中构建文化 https://korshunov.ai/pt-br/article/17971-a-ai-breakroom-permite-que-agentes-construam-cultura-em-salas-de-chat/korshunov.aipt2026-08-12T03:29:17.741368+00:00A AI Breakroom permite que agentes construam cultura em salas de chat compartilhadas https://korshunov.ai/hi/article/17971-ai-breakroom-ejentts-ko-saajhaa-caitt-ruumon-men-snskrti-bnaane-detaa-hai/korshunov.aihi2026-08-12T03:29:17.741368+00:00AI Breakroom एजेंट्स को साझा चैट रूमों में संस्कृति बनाने देता है https://korshunov.ai/ja/article/17971-ai-breakroomhae-shientokagong-you-chiyatsutoru-mutewen-hua-wogou-zhu/korshunov.aija2026-08-12T03:29:17.741368+00:00AI Breakroomはエージェントが共有チャットルームで文化を構築できるようにする https://korshunov.ai/ko/article/17971-ai-breakroomeun-eijeonteuga-gongyu-caetingbangeseo-munhwareul-gucughal-su-issge/korshunov.aiko2026-08-12T03:29:17.741368+00:00AI Breakroom은 에이전트가 공유 채팅방에서 문화를 구축할 수 있게 합니다 https://korshunov.ai/fr/article/17971-ai-breakroom-permet-aux-agents-de-construire-une-culture-dans-des-salles-de-chat/korshunov.aifr2026-08-12T03:29:17.741368+00:00AI Breakroom permet aux agents de construire une culture dans des salles de chat partagées https://korshunov.ai/id/article/17971-ai-breakroom-memungkinkan-agen-membangun-budaya-di-ruang-obrolan-bersama/korshunov.aiid2026-08-12T03:29:17.741368+00:00AI Breakroom memungkinkan agen membangun budaya di ruang obrolan bersama https://korshunov.ai/ar/article/17971-grf-strh-ldhk-lstn-y-ttyh-llwkl-bn-thqf-fy-grf-drdsh-mshtrk/korshunov.aiar2026-08-12T03:29:17.741368+00:00غرفة استراحة الذكاء الاصطناعي تتيح للوكلاء بناء ثقافة في غرف دردشة مشتركة https://korshunov.ai/en/article/17970-crewai-v1-15-15-adds-flow-outcome-reporting-and-bumps-torch/korshunov.aien2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 adds flow outcome reporting and bumps torch https://korshunov.ai/ru/article/17970-crewai-v1-15-15-dobavliaet-otchetnost-o-rezultatakh-potoka-i-obnovliaet-torch/korshunov.airu2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 добавляет отчётность о результатах потока и обновляет torch https://korshunov.ai/es/article/17970-crewai-v1-15-15-anade-informes-de-resultados-de-flujo-y-actualiza-torch/korshunov.aies2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 añade informes de resultados de flujo y actualiza torch https://korshunov.ai/zh-hans/article/17970-crewai-v1-15-15-xin-zeng-liu-cheng-jie-guo-bao-gao-bing-sheng-ji-torch/korshunov.aizh-cn2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 新增流程结果报告并升级 torch https://korshunov.ai/pt-br/article/17970-crewai-v1-15-15-adiciona-relatorios-de-resultado-de-fluxo-e-atualiza-torch/korshunov.aipt2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 adiciona relatórios de resultado de fluxo e atualiza torch https://korshunov.ai/hi/article/17970-crewai-v1-15-15-men-phlo-aauttputt-riportting-joddii-gii-aur-torch-ko-apddett/korshunov.aihi2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 में फ्लो आउटपुट रिपोर्टिंग जोड़ी गई और torch को अपडेट किया गया https://korshunov.ai/ja/article/17970-crewai-v1-15-15-kahuro-jie-guo-noreho-toto-torch-noha-shiyonatsuhuwozhui-jia/korshunov.aija2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 がフロー結果のレポートと torch のバージョンアップを追加 https://korshunov.ai/ko/article/17970-crewai-v1-15-15ga-heureum-gyeolgwa-bogowa-torch-eobgeureideureul-cugahabnida/korshunov.aiko2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15가 흐름 결과 보고와 torch 업그레이드를 추가합니다 https://korshunov.ai/fr/article/17970-crewai-v1-15-15-ajoute-le-rapport-de-resultat-de-flux-et-met-a-jour-torch/korshunov.aifr2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 ajoute le rapport de résultat de flux et met à jour torch https://korshunov.ai/id/article/17970-crewai-v1-15-15-menambahkan-pelaporan-hasil-aliran-dan-meningkatkan-torch/korshunov.aiid2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 menambahkan pelaporan hasil aliran dan meningkatkan torch https://korshunov.ai/ar/article/17970-crewai-v1-15-15-ydyf-tqryr-ntyj-ltdfq-wyzyd-mn-torch/korshunov.aiar2026-08-12T03:21:55.033822+00:00CrewAI v1.15.15 يضيف تقارير نتيجة التدفق ويزيد من torch https://korshunov.ai/en/article/17848-milmmt-46-v1-0-improves-multilingual-translation-via-reference-free-post/korshunov.aien2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 improves multilingual translation via reference-free post-training https://korshunov.ai/ru/article/17848-milmmt-46-v1-0-uluchshaet-mnogoiazychnyi-perevod-s-pomoshchiu-obucheniia-posle/korshunov.airu2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 улучшает многоязычный перевод с помощью обучения после тренировки без эталонов https://korshunov.ai/es/article/17848-milmmt-46-v1-0-mejora-la-traduccion-multilingue-mediante-postentrenamiento-sin/korshunov.aies2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 mejora la traducción multilingüe mediante postentrenamiento sin referencia https://korshunov.ai/zh-hans/article/17848-milmmt-46-v1-0-tong-guo-wu-can-kao-hou-xun-lian-ti-sheng-duo-yu-yan-fan-yi/korshunov.aizh-cn2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 通过无参考后训练提升多语言翻译 https://korshunov.ai/pt-br/article/17848-milmmt-46-v1-0-melhora-a-traducao-multilingue-por-meio-de-pos-treinamento-sem/korshunov.aipt2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 melhora a tradução multilíngue por meio de pós-treinamento sem referência https://korshunov.ai/hi/article/17848-milmmt-46-v1-0-ne-sndrbh-mukt-postt-ttrening-ke-maadhym-se-bhubhaassii-anuvaad/korshunov.aihi2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 ने संदर्भ-मुक्त पोस्ट-ट्रेनिंग के माध्यम से बहुभाषी अनुवाद में सुधार किया https://korshunov.ai/ja/article/17848-milmmt-46-v1-0-hacan-zhao-nashinohosutotore-ninkuniyoriduo-yan-yu-fan-yi-wogai/korshunov.aija2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 は参照なしのポストトレーニングにより多言語翻訳を改善 https://korshunov.ai/ko/article/17848-milmmt-46-v1-0-repeoreonseu-eobsneun-sahu-hagseubeul-tonghae-dagugeo-beonyeog/korshunov.aiko2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0, 레퍼런스 없는 사후 학습을 통해 다국어 번역 성능 향상 https://korshunov.ai/fr/article/17848-milmmt-46-v1-0-ameliore-la-traduction-multilingue-grace-a-un-post-apprentissage/korshunov.aifr2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 améliore la traduction multilingue grâce à un post-apprentissage sans référence https://korshunov.ai/id/article/17848-milmmt-46-v1-0-meningkatkan-terjemahan-multibahasa-melalui-pasca-pelatihan-tanpa/korshunov.aiid2026-08-12T03:14:41.392374+00:00MiLMMT-46-v1.0 meningkatkan terjemahan multibahasa melalui pasca-pelatihan tanpa referensi https://korshunov.ai/ar/article/17848-thswn-milmmt-46-v1-0-lltrjm-mt-dd-llgt-br-m-b-d-ltdryb-bdwn-mrj/korshunov.aiar2026-08-12T03:14:41.392374+00:00تحسّن MiLMMT-46-v1.0 للترجمة متعددة اللغات عبر ما بعد التدريب بدون مرجع https://korshunov.ai/hi/article/17816-anthropic-openai-google-meta-microsoft-aur-mistral-ne-eu-ai-paardrshitaa-kodd-pr/korshunov.aihi2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft और Mistral ने EU AI पारदर्शिता कोड पर हस्ताक्षर किए https://korshunov.ai/ko/article/17816-anthropic-openai-google-meta-microsoft-mistrali-eu-ai-tumyeongseong-kodeue/korshunov.aiko2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft, Mistral이 EU AI 투명성 코드에 서명 https://korshunov.ai/fr/article/17816-anthropic-openai-google-meta-microsoft-et-mistral-signent-le-code-de-de-l-ia-de/korshunov.aifr2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft et Mistral signent le code de transparence de l'IA de l'UE https://korshunov.ai/id/article/17816-anthropic-openai-google-meta-microsoft-dan-mistral-menandatangani-kode-ai-ue/korshunov.aiid2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft, dan Mistral menandatangani kode transparansi AI UE https://korshunov.ai/ar/article/17816-nthrwbyk-wwbn-yh-y-wjwjl-wmyt-wmykrwswft-wmstrl-twq-n-mdwn-lshffy-lkhs-bldhk-y/korshunov.aiar2026-08-12T00:47:19.122118+00:00أنثروبيك وأوبن إيه آي وجوجل وميتا ومايكروسوفت ومسترال توقعان مدونة الشفافية الخاصة بالذكاء الاصطناعي في الاتحاد الأوروبي https://korshunov.ai/en/article/17816-anthropic-openai-google-meta-microsoft-and-mistral-sign-eu-ai-transparency-code/korshunov.aien2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft, and Mistral sign EU AI transparency code https://korshunov.ai/ru/article/17816-anthropic-openai-google-meta-microsoft-i-mistral-podpisali-kodeks-prozrachnosti/korshunov.airu2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft и Mistral подписали кодекс прозрачности ИИ ЕС https://korshunov.ai/es/article/17816-anthropic-openai-google-meta-microsoft-y-mistral-firman-el-codigo-de-de-ia-de-la/korshunov.aies2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft y Mistral firman el código de transparencia de IA de la UE https://korshunov.ai/pt-br/article/17816-anthropic-openai-google-meta-microsoft-e-mistral-assinam-codigo-de-transparencia/korshunov.aipt2026-08-12T00:47:19.122118+00:00Anthropic, OpenAI, Google, Meta, Microsoft e Mistral assinam código de transparência de IA da UE https://korshunov.ai/en/article/17810-openai-makes-daybreak-models-available-on-aws-via-amazon-bedrock/korshunov.aien2026-08-11T22:09:01.669998+00:00OpenAI makes Daybreak models available on AWS via Amazon Bedrock https://korshunov.ai/ru/article/17810-openai-predostavliaet-modeli-daybreak-v-aws-cherez-amazon-bedrock/korshunov.airu2026-08-11T22:09:01.669998+00:00OpenAI предоставляет модели Daybreak в AWS через Amazon Bedrock https://korshunov.ai/es/article/17810-openai-pone-a-disposicion-los-modelos-daybreak-en-aws-a-traves-de-amazon-bedrock/korshunov.aies2026-08-11T22:09:01.669998+00:00OpenAI pone a disposición los modelos Daybreak en AWS a través de Amazon Bedrock https://korshunov.ai/zh-hans/article/17810-openai-tong-guo-amazon-bedrock-zai-aws-shang-ti-gong-daybreak-mo-xing/korshunov.aizh-cn2026-08-11T22:09:01.669998+00:00OpenAI 通过 Amazon Bedrock 在 AWS 上提供 Daybreak 模型 https://korshunov.ai/pt-br/article/17810-openai-disponibiliza-modelos-daybreak-na-aws-via-amazon-bedrock/korshunov.aipt2026-08-11T22:09:01.669998+00:00OpenAI disponibiliza modelos Daybreak na AWS via Amazon Bedrock https://korshunov.ai/hi/article/17810-openai-ne-amazon-bedrock-ke-maadhym-se-aws-pr-daybreak-monddl-uplbdh-kraae-hain/korshunov.aihi2026-08-11T22:09:01.669998+00:00OpenAI ने Amazon Bedrock के माध्यम से AWS पर Daybreak मॉडल उपलब्ध कराए हैं https://korshunov.ai/ja/article/17810-openai-amazon-bedrockjing-you-teawsnidaybreakmoteruwoti-gong/korshunov.aija2026-08-11T22:09:01.669998+00:00OpenAI、Amazon Bedrock経由でAWSにDaybreakモデルを提供 https://korshunov.ai/ko/article/17810-openai-amazon-bedrockreul-tonghae-awseseo-daybreak-model-jegong/korshunov.aiko2026-08-11T22:09:01.669998+00:00OpenAI, Amazon Bedrock를 통해 AWS에서 Daybreak 모델 제공 https://korshunov.ai/fr/article/17810-openai-rend-les-modeles-daybreak-disponibles-sur-aws-via-amazon-bedrock/korshunov.aifr2026-08-11T22:09:01.669998+00:00OpenAI rend les modèles Daybreak disponibles sur AWS via Amazon Bedrock https://korshunov.ai/id/article/17810-openai-menyediakan-model-daybreak-di-aws-melalui-amazon-bedrock/korshunov.aiid2026-08-11T22:09:01.669998+00:00OpenAI menyediakan model Daybreak di AWS melalui Amazon Bedrock https://korshunov.ai/ar/article/17810-openai-tj-l-nmdhj-daybreak-mth-l-aws-br-amazon-bedrock/korshunov.aiar2026-08-11T22:09:01.669998+00:00OpenAI تجعل نماذج Daybreak متاحة على AWS عبر Amazon Bedrock https://korshunov.ai/en/article/17803-ltx-releases-ltx-2-5-an-open-weights-world-model-for-local-video-generation/korshunov.aien2026-08-11T21:58:00.622475+00:00LTX releases LTX-2.5, an open weights world model for local video generation https://korshunov.ai/ru/article/17803-ltx-vypuskaet-ltx-2-5-model-mira-s-otkrytymi-vesami-dlia-lokalnoi-generatsii/korshunov.airu2026-08-11T21:58:00.622475+00:00LTX выпускает LTX-2.5, модель мира с открытыми весами для локальной генерации видео https://korshunov.ai/es/article/17803-ltx-lanza-ltx-2-5-un-modelo-de-mundo-con-pesos-abiertos-para-generacion-de-video/korshunov.aies2026-08-11T21:58:00.622475+00:00LTX lanza LTX-2.5, un modelo de mundo con pesos abiertos para generación de video local https://korshunov.ai/zh-hans/article/17803-ltxfa-bu-ltx-2-5-yi-kuan-yong-yu-ben-di-shi-pin-sheng-cheng-de-kai-yuan-quan-shi/korshunov.aizh-cn2026-08-11T21:58:00.622475+00:00LTX发布LTX-2.5，一款用于本地视频生成的开源权重世界模型 https://korshunov.ai/pt-br/article/17803-ltx-lanca-ltx-2-5-um-modelo-de-mundo-com-pesos-abertos-para-geracao-local-de/korshunov.aipt2026-08-11T21:58:00.622475+00:00LTX lança LTX-2.5, um modelo de mundo com pesos abertos para geração local de vídeo https://korshunov.ai/hi/article/17803-ltx-ne-sthaaniiy-viiddiyo-nirmaann-ke-lie-khule-vjn-vaale-vishv-monddl-ltx-2-5/korshunov.aihi2026-08-11T21:58:00.622475+00:00LTX ने स्थानीय वीडियो निर्माण के लिए खुले वजन वाले विश्व मॉडल LTX-2.5 को जारी किया https://korshunov.ai/ja/article/17803-ltxkaro-karudong-hua-sheng-cheng-yong-noo-hunueito-wa-rutomoteru-ltx-2-5-woriri/korshunov.aija2026-08-11T21:58:00.622475+00:00LTXがローカル動画生成用のオープンウェイト・ワールドモデル「LTX-2.5」をリリース https://korshunov.ai/ko/article/17803-ltxga-rokeol-bidio-saengseongeul-wihan-opeun-gajungci-weoldeu-model-ltx-2-5reul/korshunov.aiko2026-08-11T21:58:00.622475+00:00LTX가 로컬 비디오 생성을 위한 오픈 가중치 월드 모델 LTX-2.5를 출시 https://korshunov.ai/fr/article/17803-ltx-publie-ltx-2-5-un-modele-du-monde-a-poids-ouverts-pour-la-generation-video/korshunov.aifr2026-08-11T21:58:00.622475+00:00LTX publie LTX-2.5, un modèle du monde à poids ouverts pour la génération vidéo locale https://korshunov.ai/id/article/17803-ltx-merilis-ltx-2-5-model-dunia-bobot-terbuka-untuk-generasi-video-lokal/korshunov.aiid2026-08-11T21:58:00.622475+00:00LTX merilis LTX-2.5, model dunia bobot terbuka untuk generasi video lokal https://korshunov.ai/ar/article/17803-ltx-ttlq-ltx-2-5-nmwdhj-lmy-bwzn-mftwh-ltwlyd-lfydyw-mhlyan/korshunov.aiar2026-08-11T21:58:00.622475+00:00LTX تطلق LTX-2.5، نموذج عالمي بأوزان مفتوحة لتوليد الفيديو محلياً https://korshunov.ai/en/article/17797-microsoft-research-introduces-care-x-a-unified-chest-x-ray-vlm-with-auxiliary/korshunov.aien2026-08-11T21:31:05.159661+00:00Microsoft Research introduces CARE-X, a unified chest X-ray VLM with auxiliary supervision and tool-augmented measurement https://korshunov.ai/ru/article/17797-microsoft-research-predstavliaet-care-x-unifitsirovannuiu-vlm-dlia-grudnoi-s-i/korshunov.airu2026-08-11T21:31:05.159661+00:00Microsoft Research представляет CARE-X: унифицированную VLM для рентгенографии грудной клетки с вспомогательным обучением и измерительными инструментами https://korshunov.ai/es/article/17797-microsoft-research-presenta-care-x-un-vlm-unificado-de-radiografias-de-torax-con/korshunov.aies2026-08-11T21:31:05.159661+00:00Microsoft Research presenta CARE-X, un VLM unificado de radiografías de tórax con supervisión auxiliar y medición aumentada por herramientas https://korshunov.ai/zh-hans/article/17797-wei-ruan-yan-jiu-yuan-tui-chu-care-x-yi-chong-dai-you-fu-zhu-jian-du-he-gong-ju/korshunov.aizh-cn2026-08-11T21:31:05.159661+00:00微软研究院推出CARE-X：一种带有辅助监督和工具增强测量的统一胸部X射线VLM https://korshunov.ai/pt-br/article/17797-microsoft-research-apresenta-care-x-um-vlm-unificado-de-raios-x-toracico-com-e/korshunov.aipt2026-08-11T21:31:05.159661+00:00Microsoft Research apresenta CARE-X, um VLM unificado de raios X torácico com supervisão auxiliar e medição aumentada por ferramentas https://korshunov.ai/hi/article/17797-maaikrosonphtt-risrc-ne-care-x-pesh-kiyaa-ek-ekiikrt-chaatii-kii-eks-re-vlm-aur/korshunov.aihi2026-08-11T21:31:05.159661+00:00माइक्रोसॉफ्ट रिसर्च ने CARE-X पेश किया, एक एकीकृत छाती की एक्स-रे VLM जिसमें सहायक निगरानी और उपकरण-संवर्धित माप शामिल हैं https://korshunov.ai/ja/article/17797-microsoft-researchkabu-zhu-jiao-shi-xin-hao-totsu-rukuo-zhang-ce-ding-ji-neng-he/korshunov.aija2026-08-11T21:31:05.159661+00:00Microsoft Researchが補助教師信号とツール拡張測定機能を備えた統合胸部X線VLM「CARE-X」を発表 https://korshunov.ai/ko/article/17797-maikeurosopeuteu-riseoci-bojo-gamdoggwa-dogu-jeunggang-ceugjeongeul-gajcun-x-ray/korshunov.aiko2026-08-11T21:31:05.159661+00:00마이크로소프트 리서치, 보조 감독과 도구 증강 측정을 갖춘 통합chest X-ray VLM인 CARE-X를 선보임 https://korshunov.ai/fr/article/17797-microsoft-research-presente-care-x-un-vlm-unifie-de-radiographie-thoracique-avec/korshunov.aifr2026-08-11T21:31:05.159661+00:00Microsoft Research présente CARE-X, un VLM unifié de radiographie thoracique avec supervision auxiliaire et mesure augmentée par outils https://korshunov.ai/id/article/17797-microsoft-research-memperkenalkan-care-x-vlm-rontgen-dada-terpadu-dengan-dan/korshunov.aiid2026-08-11T21:31:05.159661+00:00Microsoft Research memperkenalkan CARE-X, VLM rontgen dada terpadu dengan supervisi tambahan dan pengukuran yang diperkaya alat https://korshunov.ai/ar/article/17797-bhth-mykrwswft-tqdm-care-x-nmwdhj-rwy-wlg-mwhd-lsh-lsdr-m-shrf-ms-d-wqys-md-wm/korshunov.aiar2026-08-11T21:31:05.159661+00:00أبحاث مايكروسوفت تقدم CARE-X، نموذج رؤية ولغة موحد لأشعة الصدر مع إشراف مساعد وقياس مدعوم بالأدوات https://korshunov.ai/zh-hans/article/17796-google-jie-zhu-amie-zhan-shi-shi-shi-lin-chuang-shi-pin-wen-zhen-neng-li/korshunov.aizh-cn2026-08-11T21:05:24.891835+00:00Google 借助 AMIE 展示实时临床视频问诊能力 https://korshunov.ai/en/article/17796-google-demonstrates-real-time-clinical-video-consultation-capabilities-with-amie/korshunov.aien2026-08-11T21:05:24.891835+00:00Google demonstrates real-time clinical video consultation capabilities with AMIE https://korshunov.ai/ru/article/17796-google-demonstriruet-vozmozhnosti-klinicheskikh-videokonsultatsii-v-realnom-s/korshunov.airu2026-08-11T21:05:24.891835+00:00Google демонстрирует возможности клинических видеоконсультаций в реальном времени с помощью AMIE https://korshunov.ai/es/article/17796-google-demuestra-capacidades-de-consulta-clinica-por-video-en-tiempo-real-con/korshunov.aies2026-08-11T21:05:24.891835+00:00Google demuestra capacidades de consulta clínica por video en tiempo real con AMIE https://korshunov.ai/pt-br/article/17796-google-demonstra-capacidades-de-consulta-clinica-por-video-em-tempo-real-com-o/korshunov.aipt2026-08-11T21:05:24.891835+00:00Google demonstra capacidades de consulta clínica por vídeo em tempo real com o AMIE https://korshunov.ai/hi/article/17796-google-ne-amie-ke-saath-riyl-ttaaim-klinikl-viiddiyo-praamrsh-kssmtaaon-kaa/korshunov.aihi2026-08-11T21:05:24.891835+00:00Google ने AMIE के साथ रियल-टाइम क्लिनिकल वीडियो परामर्श क्षमताओं का प्रदर्शन किया https://korshunov.ai/ja/article/17796-google-amiewoyong-itariarutaimulin-chuang-hiteoxiang-tan-neng-li-woshi-zheng/korshunov.aija2026-08-11T21:05:24.891835+00:00Google、AMIEを用いたリアルタイム臨床ビデオ相談能力を実証 https://korshunov.ai/ko/article/17796-gugeul-amiero-silsigan-imsang-bidio-sangdam-gineung-siyeon/korshunov.aiko2026-08-11T21:05:24.891835+00:00구글, AMIE로 실시간 임상 비디오 상담 기능 시연 https://korshunov.ai/fr/article/17796-google-demontre-des-capacites-de-consultation-video-clinique-en-temps-reel-avec/korshunov.aifr2026-08-11T21:05:24.891835+00:00Google démontre des capacités de consultation vidéo clinique en temps réel avec AMIE https://korshunov.ai/id/article/17796-google-mendemonstrasikan-kemampuan-konsultasi-video-klinis-waktu-nyata-dengan/korshunov.aiid2026-08-11T21:05:24.891835+00:00Google mendemonstrasikan kemampuan konsultasi video klinis waktu nyata dengan AMIE https://korshunov.ai/ar/article/17796-jwjl-tuzhr-qdrt-lstshr-lsryry-blfydyw-fy-lwqt-lf-ly-bstkhdm-amie/korshunov.aiar2026-08-11T21:05:24.891835+00:00جوجل تُظهر قدرات الاستشارة السريرية بالفيديو في الوقت الفعلي باستخدام AMIE https://korshunov.ai/en/article/17794-claude-officially-marks-ai-generated-content-steganographically/korshunov.aien2026-08-11T20:20:51.923502+00:00Claude officially marks AI-generated content steganographically https://korshunov.ai/ru/article/17794-claude-ofitsialno-markiruet-ii-kontent-steganograficheski/korshunov.airu2026-08-11T20:20:51.923502+00:00Claude официально маркирует ИИ-контент стеганографически https://korshunov.ai/es/article/17794-claude-marca-oficialmente-el-contenido-generado-por-ia-esteganograficamente/korshunov.aies2026-08-11T20:20:51.923502+00:00Claude marca oficialmente el contenido generado por IA esteganográficamente https://korshunov.ai/pt-br/article/17794-claude-marca-oficialmente-conteudo-gerado-por-ia-estenograficamente/korshunov.aipt2026-08-11T20:20:51.923502+00:00Claude marca oficialmente conteúdo gerado por IA estenograficamente https://korshunov.ai/hi/article/17794-claude-aadhikaarik-taur-pr-ai-jnit-saamgrii-ko-sttegnograaphik-ruup-se-cihnit/korshunov.aihi2026-08-11T20:20:51.923502+00:00Claude आधिकारिक तौर पर AI-जनित सामग्री को स्टेगनोग्राफिक रूप से चिह्नित करता है https://korshunov.ai/ja/article/17794-claudekaaisheng-cheng-kontentsuwosutekanokurahuide-nizheng-shi-nima-kinku/korshunov.aija2026-08-11T20:20:51.923502+00:00ClaudeがAI生成コンテンツをステガノグラフィ的に正式にマーキング https://korshunov.ai/ko/article/17794-claudega-ai-saengseong-kontenceureul-seuteganogeuraepiro-gongsigjeogeuro-pyosi/korshunov.aiko2026-08-11T20:20:51.923502+00:00Claude가 AI 생성 콘텐츠를 스테가노그래피로 공식적으로 표시 https://korshunov.ai/fr/article/17794-claude-marque-officiellement-le-contenu-genere-par-l-ia-de-maniere/korshunov.aifr2026-08-11T20:20:51.923502+00:00Claude marque officiellement le contenu généré par l'IA de manière stéganographique https://korshunov.ai/id/article/17794-claude-secara-resmi-menandai-konten-yang-dihasilkan-ai-secara-steganografis/korshunov.aiid2026-08-11T20:20:51.923502+00:00Claude secara resmi menandai konten yang dihasilkan AI secara steganografis https://korshunov.ai/ar/article/17794-klwd-y-lwn-rsmyan-n-wd-lm-l-lmhtw-lmuwlwd-bldhk-lstn-y-bshkl-stygnwgrfy/korshunov.aiar2026-08-11T20:20:51.923502+00:00كلاود يعلّن رسمياً عن وضع علامة على المحتوى المُولّد بالذكاء الاصطناعي بشكل ستيغانوغرافي https://korshunov.ai/ar/article/17793-slht-lsdr-2-1-228-mn-claude-code-lktshf-git-wm-lj-bynt-l-tmd-wkht-ljls/korshunov.aiar2026-08-11T20:16:16.797320+00:00إصلاحات الإصدار 2.1.228 من Claude Code لاكتشاف Git، ومعالجة بيانات الاعتماد، وأخطاء الجلسة https://korshunov.ai/en/article/17793-claude-code-v2-1-228-fixes-git-detection-credential-handling-and-session-bugs/korshunov.aien2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 fixes git detection, credential handling, and session bugs https://korshunov.ai/ru/article/17793-claude-code-v2-1-228-ispravliaet-obnaruzhenie-git-obrabotku-uchetnykh-dannykh-i/korshunov.airu2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 исправляет обнаружение git, обработку учётных данных и ошибки сессий https://korshunov.ai/es/article/17793-claude-code-v2-1-228-corrige-la-deteccion-de-git-el-manejo-de-credenciales-y-de/korshunov.aies2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 corrige la detección de git, el manejo de credenciales y errores de sesión https://korshunov.ai/zh-hans/article/17793-claude-code-v2-1-228-xiu-fu-liao-git-jian-ce-ping-ju-chu-li-he-hui-hua-cuo-wu/korshunov.aizh-cn2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 修复了 git 检测、凭据处理和会话错误 https://korshunov.ai/pt-br/article/17793-claude-code-v2-1-228-corrige-deteccao-de-git-manipulacao-de-credenciais-e-bugs/korshunov.aipt2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 corrige detecção de git, manipulação de credenciais e bugs de sessão https://korshunov.ai/hi/article/17793-claude-code-v2-1-228-ne-git-phcaan-kreddenshiyl-hainddling-aur-str-bg-tthii-kie/korshunov.aihi2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 ने git पहचान, क्रेडेंशियल हैंडलिंग और सत्र बग ठी किए https://korshunov.ai/ja/article/17793-claude-code-v2-1-228-ka-git-nojian-chu-ren-zheng-qing-bao-nochu-li-zheng/korshunov.aija2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 が git の検出、認証情報の処理、セッションのバグを修正 https://korshunov.ai/ko/article/17793-claude-code-v2-1-228i-git-gamji-jagyeog-jeungmyeong-ceori-mic-sesyeon-beogeureul/korshunov.aiko2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228이 Git 감지, 자격 증명 처리 및 세션 버그를 수정했습니다 https://korshunov.ai/fr/article/17793-claude-code-v2-1-228-corrige-la-detection-git-la-gestion-des-identifiants-et-les/korshunov.aifr2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 corrige la détection git, la gestion des identifiants et les bugs de session https://korshunov.ai/id/article/17793-claude-code-v2-1-228-memperbaiki-deteksi-git-penanganan-kredensial-dan-bug-sesi/korshunov.aiid2026-08-11T20:16:16.797320+00:00Claude Code v2.1.228 memperbaiki deteksi git, penanganan kredensial, dan bug sesi https://korshunov.ai/en/article/17790-xai-launches-grok-bot-always-on-ai-agents-for-desktop-and-ios/korshunov.aien2026-08-11T19:06:39.483122+00:00xAI launches Grok Bot, always-on AI agents for desktop and iOS https://korshunov.ai/ru/article/17790-xai-zapuskaet-grok-bot-vsegda-aktivnye-ii-agenty-dlia-desktopa-i-ios/korshunov.airu2026-08-11T19:06:39.483122+00:00xAI запускает Grok Bot: всегда активные ИИ-агенты для десктопа и iOS https://korshunov.ai/es/article/17790-xai-lanza-grok-bot-agentes-de-ia-siempre-activos-para-escritorio-e-ios/korshunov.aies2026-08-11T19:06:39.483122+00:00xAI lanza Grok Bot, agentes de IA siempre activos para escritorio e iOS https://korshunov.ai/zh-hans/article/17790-xai-tui-chu-grok-bot-mian-xiang-zhuo-mian-he-ios-de-chang-zhu-ai-zhi-neng-ti/korshunov.aizh-cn2026-08-11T19:06:39.483122+00:00xAI 推出 Grok Bot，面向桌面和 iOS 的常驻 AI 智能体 https://korshunov.ai/pt-br/article/17790-xai-lanca-grok-bot-agentes-de-ia-sempre-ativos-para-desktop-e-ios/korshunov.aipt2026-08-11T19:06:39.483122+00:00xAI lança Grok Bot, agentes de IA sempre ativos para desktop e iOS https://korshunov.ai/hi/article/17790-xai-ne-ddeskttonp-aur-ios-ke-lie-hmeshaa-skriy-ai-ejentts-grok-bot-lonnc-kie/korshunov.aihi2026-08-11T19:06:39.483122+00:00xAI ने डेस्कटॉप और iOS के लिए हमेशा-सक्रिय AI एजेंट्स Grok Bot लॉन्च किए https://korshunov.ai/ja/article/17790-xaikagrok-botworiri-su-tesukutotsuhuoyohiiosxiang-kenochang-shi-onaie-shiento/korshunov.aija2026-08-11T19:06:39.483122+00:00xAIがGrok Botをリリース、デスクトップおよびiOS向けの常時オンAIエージェント https://korshunov.ai/ko/article/17790-xai-deseukeutob-mic-iosyong-sangsi-hwalseong-ai-eijeonteu-grok-bot-culsi/korshunov.aiko2026-08-11T19:06:39.483122+00:00xAI, 데스크톱 및 iOS용 상시 활성 AI 에이전트 Grok Bot 출시 https://korshunov.ai/fr/article/17790-xai-lance-grok-bot-des-agents-ia-toujours-actifs-pour-desktop-et-ios/korshunov.aifr2026-08-11T19:06:39.483122+00:00xAI lance Grok Bot, des agents IA toujours actifs pour desktop et iOS https://korshunov.ai/id/article/17790-xai-meluncurkan-grok-bot-agen-ai-yang-selalu-aktif-untuk-desktop-dan-ios/korshunov.aiid2026-08-11T19:06:39.483122+00:00xAI meluncurkan Grok Bot, agen AI yang selalu aktif untuk desktop dan iOS https://korshunov.ai/ar/article/17790-ks-y-ttlq-grok-bot-wkl-dhk-stn-y-y-mlwn-dyman-l-sth-lmktb-wios/korshunov.aiar2026-08-11T19:06:39.483122+00:00إكس آي تطلق Grok Bot، وكلاء ذكاء اصطناعي يعملون دائمًا على سطح المكتب وiOS https://korshunov.ai/en/article/17788-llama-cpp-b10361-fixes-swa-for-exaone-4-5-and-skips-tensor-skip-in-metadata-path/korshunov.aien2026-08-11T18:22:51.195080+00:00llama.cpp b10361 fixes SWA for EXAONE 4.5 and skips TENSOR_SKIP in metadata path https://korshunov.ai/ko/article/17788-llama-cpp-b10361i-exaone-4-5yi-swareul-sujeonghago-metadeiteo-gyeongroeseo/korshunov.aiko2026-08-11T18:22:51.195080+00:00llama.cpp b10361이 EXAONE 4.5의 SWA를 수정하고 메타데이터 경로에서 TENSOR_SKIP을 건너뜀 https://korshunov.ai/ru/article/17788-llama-cpp-b10361-ispravliaet-swa-dlia-exaone-4-5-i-propuskaet-tensor-skip-v-puti/korshunov.airu2026-08-11T18:22:51.195080+00:00llama.cpp b10361 исправляет SWA для EXAONE 4.5 и пропускает TENSOR_SKIP в пути метаданных https://korshunov.ai/es/article/17788-llama-cpp-b10361-corrige-swa-para-exaone-4-5-y-omite-tensor-skip-en-la-ruta-de/korshunov.aies2026-08-11T18:22:51.195080+00:00llama.cpp b10361 corrige SWA para EXAONE 4.5 y omite TENSOR_SKIP en la ruta de metadatos https://korshunov.ai/zh-hans/article/17788-llama-cpp-b10361-xiu-fu-exaone-4-5-de-swa-bing-tiao-guo-yuan-shu-ju-lu-jing-de/korshunov.aizh-cn2026-08-11T18:22:51.195080+00:00llama.cpp b10361 修复 EXAONE 4.5 的 SWA 并跳过元数据路径中的 TENSOR_SKIP https://korshunov.ai/pt-br/article/17788-llama-cpp-b10361-corrige-swa-para-exaone-4-5-e-ignora-tensor-skip-no-caminho-de/korshunov.aipt2026-08-11T18:22:51.195080+00:00llama.cpp b10361 corrige SWA para EXAONE 4.5 e ignora TENSOR_SKIP no caminho de metadados https://korshunov.ai/hi/article/17788-llama-cpp-b10361-ne-exaone-4-5-ke-lie-swa-ko-tthiik-kiyaa-aur-mettaaddettaa-pth/korshunov.aihi2026-08-11T18:22:51.195080+00:00llama.cpp b10361 ने EXAONE 4.5 के लिए SWA को ठीक किया और मेटाडेटा पथ में TENSOR_SKIP को छोड़ा https://korshunov.ai/ja/article/17788-llama-cpp-b10361teexaone-4-5noswawoxiu-zheng-shi-metate-tahasutenotensor/korshunov.aija2026-08-11T18:22:51.195080+00:00llama.cpp b10361でEXAONE 4.5のSWAを修正し、メタデータパスでのTENSOR_SKIPをスキップ https://korshunov.ai/fr/article/17788-llama-cpp-b10361-corrige-swa-pour-exaone-4-5-et-ignore-tensor-skip-dans-le-des/korshunov.aifr2026-08-11T18:22:51.195080+00:00llama.cpp b10361 corrige SWA pour EXAONE 4.5 et ignore TENSOR_SKIP dans le chemin des métadonnées https://korshunov.ai/id/article/17788-llama-cpp-b10361-memperbaiki-swa-untuk-exaone-4-5-dan-melewatkan-tensor-skip-di/korshunov.aiid2026-08-11T18:22:51.195080+00:00llama.cpp b10361 memperbaiki SWA untuk EXAONE 4.5 dan melewatkan TENSOR_SKIP di jalur metadata https://korshunov.ai/ar/article/17788-slh-llama-cpp-b10361-l-swa-lkhs-b-exaone-4-5-wtkhty-tensor-skip-fy-msr-lbynt/korshunov.aiar2026-08-11T18:22:51.195080+00:00إصلاح llama.cpp b10361 لـ SWA الخاص بـ EXAONE 4.5 وتخطي TENSOR_SKIP في مسار البيانات الوصفية https://korshunov.ai/en/article/17784-mistral-launches-regional-endpoints-open-model-support-and-european-compute/korshunov.aien2026-08-11T17:03:43.667223+00:00Mistral launches regional endpoints, open model support, and European compute coalition https://korshunov.ai/ru/article/17784-mistral-zapuskaet-regionalnye-konechnye-tochki-podderzhku-otkrytykh-modelei-i/korshunov.airu2026-08-11T17:03:43.667223+00:00Mistral запускает региональные конечные точки, поддержку открытых моделей и европейскую коалицию вычислений https://korshunov.ai/es/article/17784-mistral-lanza-puntos-de-conexion-regionales-soporte-para-modelos-abiertos-y-de/korshunov.aies2026-08-11T17:03:43.667223+00:00Mistral lanza puntos de conexión regionales, soporte para modelos abiertos y coalición europea de computación https://korshunov.ai/zh-hans/article/17784-mistraltui-chu-qu-yu-duan-dian-kai-fang-mo-xing-zhi-chi-he-ou-zhou-ji-suan-lian/korshunov.aizh-cn2026-08-11T17:03:43.667223+00:00Mistral推出区域端点、开放模型支持和欧洲计算联盟 https://korshunov.ai/pt-br/article/17784-mistral-lanca-endpoints-regionais-suporte-a-modelos-abertos-e-coalizao-europeia/korshunov.aipt2026-08-11T17:03:43.667223+00:00Mistral lança endpoints regionais, suporte a modelos abertos e coalizão europeia de computação https://korshunov.ai/hi/article/17784-misttrl-ne-kssetriiy-enddponintts-opn-monddl-sportt-aur-yuuropiiy-knpyuutt-lonnc/korshunov.aihi2026-08-11T17:03:43.667223+00:00मिस्ट्रल ने क्षेत्रीय एंडपॉइंट्स, ओपन मॉडल सपोर्ट और यूरोपीय कंप्यूट कोएलिशन लॉन्च किए https://korshunov.ai/ja/article/17784-mistralkadi-yu-entohointo-o-hunmoterudui-ying-ou-zhou-noji-suan-lian-he-woli-ke/korshunov.aija2026-08-11T17:03:43.667223+00:00Mistralが地域エンドポイント、オープンモデル対応、欧州の計算連合を立ち上げ https://korshunov.ai/ko/article/17784-mistral-jiyeog-endeupointeuwa-opeun-model-jiweon-yureob-keompyuting-yeonhab/korshunov.aiko2026-08-11T17:03:43.667223+00:00Mistral, 지역 엔드포인트와 오픈 모델 지원, 유럽 컴퓨팅 연합 출시 https://korshunov.ai/fr/article/17784-mistral-lance-des-points-de-terminaux-regionaux-le-support-des-modeles-ouverts/korshunov.aifr2026-08-11T17:03:43.667223+00:00Mistral lance des points de terminaux régionaux, le support des modèles ouverts et une coalition européenne pour la puissance de calcul https://korshunov.ai/id/article/17784-mistral-meluncurkan-endpoint-regional-dukungan-model-terbuka-dan-koalisi-eropa/korshunov.aiid2026-08-11T17:03:43.667223+00:00Mistral meluncurkan endpoint regional, dukungan model terbuka, dan koalisi komputasi Eropa https://korshunov.ai/ar/article/17784-mstrl-tutlq-nqt-nhy-qlymy-wd-m-lnmdhj-lmftwh-wthlf-lhwsb-lwrwby/korshunov.aiar2026-08-11T17:03:43.667223+00:00مسترال تُطلق نقاط نهاية إقليمية، ودعم النماذج المفتوحة، وتحالف الحوسبة الأوروبي https://korshunov.ai/es/article/17780-ibm-research-lanza-el-solucionador-neuronal-unificado-genco-para-analisis-de/korshunov.aies2026-08-11T16:15:53.724651+00:00IBM Research lanza el solucionador neuronal unificado GENCO para análisis de redes eléctricas https://korshunov.ai/hi/article/17780-paavr-gridd-vishlessnn-ke-lie-ibm-risrc-ne-genco-ektriikrt-nyuurl-sonlvr-jaarii/korshunov.aihi2026-08-11T16:15:53.724651+00:00पावर ग्रिड विश्लेषण के लिए IBM रिसर्च ने GENCO एकत्रीकृत न्यूरल सॉल्वर जारी किया https://korshunov.ai/en/article/17780-ibm-research-releases-genco-unified-neural-solver-for-power-grid-analysis/korshunov.aien2026-08-11T16:15:53.724651+00:00IBM Research releases GENCO unified neural solver for power grid analysis https://korshunov.ai/ru/article/17780-ibm-research-vypuskaet-unifitsirovannyi-neironnyi-reshatel-genco-dlia-analiza/korshunov.airu2026-08-11T16:15:53.724651+00:00IBM Research выпускает унифицированный нейронный решатель GENCO для анализа энергосистем https://korshunov.ai/zh-hans/article/17780-ibm-research-fa-bu-yong-yu-dian-wang-fen-xi-de-genco-tong-yi-shen-jing-qiu-jie/korshunov.aizh-cn2026-08-11T16:15:53.724651+00:00IBM Research 发布用于电网分析的 GENCO 统一神经求解器 https://korshunov.ai/pt-br/article/17780-ibm-research-lanca-o-solucionador-neural-unificado-genco-para-analise-de-redes/korshunov.aipt2026-08-11T16:15:53.724651+00:00IBM Research lança o solucionador neural unificado GENCO para análise de redes elétricas https://korshunov.ai/ja/article/17780-ibm-research-dian-li-xi-tong-jie-xi-yong-nogencotong-yi-niyu-rarusoruha-wogong/korshunov.aija2026-08-11T16:15:53.724651+00:00IBM Research、電力系統解析用のGENCO統一ニューラルソルバーを公開 https://korshunov.ai/ko/article/17780-ibm-research-jeonryeogmang-bunseogeul-wihan-genco-tonghab-singyeong-solbeo/korshunov.aiko2026-08-11T16:15:53.724651+00:00IBM Research, 전력망 분석을 위한 GENCO 통합 신경 솔버 공개 https://korshunov.ai/fr/article/17780-ibm-research-publie-genco-un-solveur-neural-unifie-pour-l-analyse-des-reseaux/korshunov.aifr2026-08-11T16:15:53.724651+00:00IBM Research publie GENCO, un solveur neural unifié pour l'analyse des réseaux électriques https://korshunov.ai/id/article/17780-ibm-research-merilis-genco-pemecah-neural-terpadu-untuk-analisis-jaringan/korshunov.aiid2026-08-11T16:15:53.724651+00:00IBM Research merilis GENCO, pemecah neural terpadu untuk analisis jaringan listrik https://korshunov.ai/ar/article/17780-bhth-y-by-m-ttlq-genco-hllan-swyan-mwhdan-lthlyl-shbkt-ltq/korshunov.aiar2026-08-11T16:15:53.724651+00:00أبحاث آي بي إم تطلق GENCO، حلالاً عصوياً موحداً لتحليل شبكات الطاقة https://korshunov.ai/ru/article/17775-claude-ot-anthropic-uluchshaet-otsenku-gipotezy-rimana-meta-vypuskaet-muse-gpt-5/korshunov.airu2026-08-11T15:04:22.611151+00:00Claude от Anthropic улучшает оценку гипотезы Римана; Meta выпускает Muse Glimmer; OpenAI запускает GPT-5.6-Cyber https://korshunov.ai/en/article/17775-anthropic-s-claude-improves-riemann-hypothesis-bound-meta-releases-muse-glimmer/korshunov.aien2026-08-11T15:04:22.611151+00:00Anthropic's Claude improves Riemann hypothesis bound; Meta releases Muse Glimmer; OpenAI launches GPT-5.6-Cyber https://korshunov.ai/es/article/17775-claude-de-anthropic-mejora-el-limite-de-la-hipotesis-de-riemann-meta-lanza-muse/korshunov.aies2026-08-11T15:04:22.611151+00:00Claude de Anthropic mejora el límite de la hipótesis de Riemann; Meta lanza Muse Glimmer; OpenAI presenta GPT-5.6-Cyber https://korshunov.ai/zh-hans/article/17775-anthropicde-claudegai-jin-liao-li-man-cai-xiang-xia-jie-metafa-bu-muse-glimmer-5/korshunov.aizh-cn2026-08-11T15:04:22.611151+00:00Anthropic的Claude改进了黎曼猜想下界；Meta发布Muse Glimmer；OpenAI推出GPT-5.6-Cyber https://korshunov.ai/pt-br/article/17775-claude-da-anthropic-melhora-limite-da-hipotese-de-riemann-meta-lanca-muse-openai/korshunov.aipt2026-08-11T15:04:22.611151+00:00Claude da Anthropic melhora limite da hipótese de Riemann; Meta lança Muse Glimmer; OpenAI lança GPT-5.6-Cyber https://korshunov.ai/hi/article/17775-anthropic-kaa-claude-riimain-priklpnaa-kii-siimaa-men-sudhaar-krtaa-hai-meta-ne/korshunov.aihi2026-08-11T15:04:22.611151+00:00Anthropic का Claude रीमैन परिकल्पना की सीमा में सुधार करता है; Meta ने Muse Glimmer जारी किया; OpenAI ने GPT-5.6-Cyber लॉन्च किया https://korshunov.ai/ja/article/17775-anthropicnoclaudekari-manyu-xiang-nojing-jie-wogai-shan-metakamuse-glimmerworiri/korshunov.aija2026-08-11T15:04:22.611151+00:00AnthropicのClaudeがリーマン予想の境界を改善、MetaがMuse Glimmerをリリース、OpenAIがGPT-5.6-Cyberを発表 https://korshunov.ai/ko/article/17775-anthropicyi-claudega-riman-gaseol-gyeonggye-gaeseon-metaga-muse-glimmer-culsi-5/korshunov.aiko2026-08-11T15:04:22.611151+00:00Anthropic의 Claude가 리만 가설 경계 개선; Meta가 Muse Glimmer 출시; OpenAI가 GPT-5.6-Cyber 발표 https://korshunov.ai/fr/article/17775-claude-d-anthropic-ameliore-la-borne-de-l-hypothese-de-riemann-meta-publie-muse/korshunov.aifr2026-08-11T15:04:22.611151+00:00Claude d'Anthropic améliore la borne de l'hypothèse de Riemann ; Meta publie Muse Glimmer ; OpenAI lance GPT-5.6-Cyber https://korshunov.ai/id/article/17775-claude-dari-anthropic-tingkatkan-batas-hipotesis-riemann-meta-rilis-muse-glimmer/korshunov.aiid2026-08-11T15:04:22.611151+00:00Claude dari Anthropic tingkatkan batas hipotesis Riemann; Meta rilis Muse Glimmer; OpenAI luncurkan GPT-5.6-Cyber https://korshunov.ai/ar/article/17775-thsyn-klwd-mn-nthrwbyk-lhd-frdy-rymn-myt-ttlq-mwz-jlymr-wbn-yh-y-ttrh-jy-by-ty-5/korshunov.aiar2026-08-11T15:04:22.611151+00:00تحسين كلاود من أنثروبيك لحد فرضية ريمان؛ ميتا تطلق موز جليمر؛ أوبن إيه آي تطرح جي بي تي-5.6-سايبر https://korshunov.ai/zh-hans/article/17767-altk-evolve-yi-ji-di-de-tui-li-cheng-ben-shi-xian-yu-ace-xiang-dang-huo-geng-you/korshunov.aizh-cn2026-08-11T14:33:02.613181+00:00ALTK-Evolve 以极低的推理成本实现与 ACE 相当或更优的准确率 https://korshunov.ai/en/article/17767-altk-evolve-achieves-same-or-better-accuracy-than-ace-at-a-fraction-of-inference/korshunov.aien2026-08-11T14:33:02.613181+00:00ALTK-Evolve achieves same-or-better accuracy than ACE at a fraction of inference cost https://korshunov.ai/ru/article/17767-altk-evolve-dostigaet-tochnosti-ne-khuzhe-ace-pri-znachitelno-menshei-stoimosti/korshunov.airu2026-08-11T14:33:02.613181+00:00ALTK-Evolve достигает точности не хуже ACE при значительно меньшей стоимости вывода https://korshunov.ai/es/article/17767-altk-evolve-logra-una-precision-igual-o-superior-a-la-de-ace-con-una-fraccion-de/korshunov.aies2026-08-11T14:33:02.613181+00:00ALTK-Evolve logra una precisión igual o superior a la de ACE con una fracción del costo de inferencia https://korshunov.ai/pt-br/article/17767-altk-evolve-alcanca-precisao-igual-ou-superior-a-do-ace-com-uma-fracao-do-custo/korshunov.aipt2026-08-11T14:33:02.613181+00:00ALTK-Evolve alcança precisão igual ou superior à do ACE com uma fração do custo de inferência https://korshunov.ai/hi/article/17767-altk-evolve-kaa-nisskriytaa-laagt-ke-ansh-pr-ace-ke-smaan-yaa-behtr-sttiiktaa/korshunov.aihi2026-08-11T14:33:02.613181+00:00ALTK-Evolve का निष्क्रियता लागत के अंश पर ACE के समान या बेहतर सटीकता प्राप्त होती है https://korshunov.ai/ja/article/17767-altk-evolvehatui-lun-kosutonoda-fu-xue-jian-teacetotong-deng-yi-shang-nojing-du/korshunov.aija2026-08-11T14:33:02.613181+00:00ALTK-Evolveは推論コストの大幅削減でACEと同等以上の精度を達成 https://korshunov.ai/ko/article/17767-altk-evolveneun-curon-biyongeul-keuge-julimyeonseo-acewa-dongdeunghageona-deo/korshunov.aiko2026-08-11T14:33:02.613181+00:00ALTK-Evolve는 추론 비용을 크게 줄이면서 ACE와 동등하거나 더 높은 정확도를 달성 https://korshunov.ai/fr/article/17767-altk-evolve-atteint-une-precision-egale-ou-superieure-a-celle-d-ace-pour-une-du/korshunov.aifr2026-08-11T14:33:02.613181+00:00ALTK-Evolve atteint une précision égale ou supérieure à celle d'ACE pour une fraction du coût d'inférence https://korshunov.ai/id/article/17767-altk-evolve-mencapai-akurasi-yang-sama-atau-lebih-baik-daripada-ace-dengan-biaya/korshunov.aiid2026-08-11T14:33:02.613181+00:00ALTK-Evolve mencapai akurasi yang sama atau lebih baik daripada ACE dengan biaya inferensi jauh lebih rendah https://korshunov.ai/ar/article/17767-yhqq-altk-evolve-dq-mswy-w-fdl-mn-ace-bksr-mn-tklf-lstdll/korshunov.aiar2026-08-11T14:33:02.613181+00:00يحقق ALTK-Evolve دقة مساوية أو أفضل من ACE بكسر من تكلفة الاستدلال https://korshunov.ai/en/article/17762-meta-argues-for-releasing-more-open-weight-models-and-invites-governments-to-ai/korshunov.aien2026-08-11T12:01:21.009676+00:00Meta argues for releasing more open-weight models and invites governments to test AI safety https://korshunov.ai/ru/article/17762-meta-vystupaet-za-vypusk-bolshego-kolichestva-modelei-s-otkrytym-vesom-i-ii/korshunov.airu2026-08-11T12:01:21.009676+00:00Meta выступает за выпуск большего количества моделей с открытым весом и приглашает правительства тестировать безопасность ИИ https://korshunov.ai/es/article/17762-meta-aboga-por-liberar-mas-modelos-de-peso-abierto-e-invita-a-los-gobiernos-a-la/korshunov.aies2026-08-11T12:01:21.009676+00:00Meta aboga por liberar más modelos de peso abierto e invita a los gobiernos a probar la seguridad de la IA https://korshunov.ai/pt-br/article/17762-meta-defende-lancamento-de-mais-modelos-de-peso-aberto-e-convida-governos-a-da/korshunov.aipt2026-08-11T12:01:21.009676+00:00Meta defende lançamento de mais modelos de peso aberto e convida governos a testar segurança da IA https://korshunov.ai/hi/article/17762-mettaa-adhik-opn-vett-monddl-jaarii-krne-kaa-pkss-men-hai-aur-srkaaron-ko-eaaii/korshunov.aihi2026-08-11T12:01:21.009676+00:00मेटा अधिक ओपन-वेट मॉडल जारी करने का पक्ष में है और सरकारों को एआई सुरक्षा की जांच के लिए आमंत्रित करता है https://korshunov.ai/fr/article/17762-meta-plaide-pour-la-publication-de-davantage-de-modeles-a-poids-ouverts-et-les-a/korshunov.aifr2026-08-11T12:01:21.009676+00:00Meta plaide pour la publication de davantage de modèles à poids ouverts et invite les gouvernements à tester la sécurité de l'IA https://korshunov.ai/id/article/17762-meta-berargumen-untuk-merilis-lebih-banyak-model-bobot-terbuka-dan-mengundang-ai/korshunov.aiid2026-08-11T12:01:21.009676+00:00Meta berargumen untuk merilis lebih banyak model bobot terbuka dan mengundang pemerintah untuk menguji keamanan AI https://korshunov.ai/ar/article/17762-myt-td-w-l-tlq-lmzyd-mn-lnmdhj-dht-lwzn-lmftwh-wtd-w-lhkwmt-lkhtbr-dhk-stn-y-mn/korshunov.aiar2026-08-11T12:01:21.009676+00:00ميتا تدعو إلى إطلاق المزيد من النماذج ذات الأوزان المفتوحة وتدعو الحكومات لاختبار ذكاء اصطناعي آمن https://korshunov.ai/en/article/17755-masters-distills-vision-language-models-via-masking-teacher-and-reinforcing/korshunov.aien2026-08-11T11:01:52.841769+00:00Masters distills vision-language models via masking teacher and reinforcing student https://korshunov.ai/ru/article/17755-masters-distilliruet-zritelno-iazykovye-modeli-cherez-maskirovanie-uchitelia-i/korshunov.airu2026-08-11T11:01:52.841769+00:00Masters дистиллирует зрительно-языковые модели через маскирование учителя и подкрепление студента https://korshunov.ai/es/article/17755-masters-destila-modelos-de-vision-y-lenguaje-mediante-enmascaramiento-del-y-del/korshunov.aies2026-08-11T11:01:52.841769+00:00Masters destila modelos de visión y lenguaje mediante enmascaramiento del maestro y refuerzo del estudiante https://korshunov.ai/zh-hans/article/17755-masterstong-guo-yan-ma-jiao-shi-yu-qiang-hua-xue-sheng-zheng-liu-shi-jue-yu-yan/korshunov.aizh-cn2026-08-11T11:01:52.841769+00:00Masters通过掩码教师与强化学生蒸馏视觉-语言模型 https://korshunov.ai/pt-br/article/17755-masters-destila-modelos-de-visao-e-linguagem-atraves-de-mascaramento-do-e-do/korshunov.aipt2026-08-11T11:01:52.841769+00:00Masters destila modelos de visão e linguagem através de mascaramento do professor e reforço do aluno https://korshunov.ai/hi/article/17755-masters-maasking-ttiicr-aur-sttuuddentt-ko-riinphorsing-ke-jrie-vijn-laingvej-ko/korshunov.aihi2026-08-11T11:01:52.841769+00:00Masters मास्किंग टीचर और स्टूडेंट को रिइन्फोर्सिंग के जरिए विजन-लैंग्वेज मॉडल्स को डिस्टिल करता है https://korshunov.ai/ja/article/17755-mastershamasukinkujiao-shi-toqiang-hua-xue-sheng-niyorihishiyonyan-yu-liu-suru/korshunov.aija2026-08-11T11:01:52.841769+00:00Mastersはマスキング教師と強化学生によりビジョン言語モデルを蒸留する https://korshunov.ai/ko/article/17755-mastersneun-maseuking-gyosawa-ganghwa-hagsaengeul-tonghae-bijeon-eoneo-modeleul/korshunov.aiko2026-08-11T11:01:52.841769+00:00Masters는 마스킹 교사와 강화 학생을 통해 비전-언어 모델을 증류합니다 https://korshunov.ai/fr/article/17755-masters-distille-les-modeles-vision-langage-via-le-masquage-du-professeur-et-le/korshunov.aifr2026-08-11T11:01:52.841769+00:00Masters distille les modèles vision-langage via le masquage du professeur et le renforcement de l'étudiant https://korshunov.ai/id/article/17755-masters-mendistilasi-model-visi-bahasa-melalui-guru-masker-dan-penguatan-siswa/korshunov.aiid2026-08-11T11:01:52.841769+00:00Masters mendistilasi model visi-bahasa melalui guru masker dan penguatan siswa https://korshunov.ai/ar/article/17755-masters-yukthwf-nmdhj-lrwy-wllg-br-qn-lm-lm-wt-zyz-ltlb/korshunov.aiar2026-08-11T11:01:52.841769+00:00Masters يُكثّف نماذج الرؤية واللغة عبر قناع المعلم وتعزيز الطالب https://korshunov.ai/zh-hans/article/17754-dstp-huan-jie-fu-za-mllm-tui-li-zhong-shi-jue-ling-pai-jian-zhi-de-shi-bai/korshunov.aizh-cn2026-08-11T10:45:00.017168+00:00DSTP 缓解复杂 MLLM 推理中视觉令牌剪枝的失败 https://korshunov.ai/en/article/17754-dstp-mitigates-visual-token-pruning-failures-in-complex-mllm-reasoning/korshunov.aien2026-08-11T10:45:00.017168+00:00DSTP mitigates visual token pruning failures in complex MLLM reasoning https://korshunov.ai/ru/article/17754-dstp-ustraniaet-sboi-pri-otsechenii-vizualnykh-tokenov-v-slozhnykh-zadachakh/korshunov.airu2026-08-11T10:45:00.017168+00:00DSTP устраняет сбои при отсечении визуальных токенов в сложных задачах рассуждения MLLM https://korshunov.ai/es/article/17754-dstp-mitiga-fallos-en-la-poda-de-tokens-visuales-en-el-razonamiento-complejo-de/korshunov.aies2026-08-11T10:45:00.017168+00:00DSTP mitiga fallos en la poda de tokens visuales en el razonamiento complejo de MLLM https://korshunov.ai/pt-br/article/17754-dstp-mitiga-falhas-na-poda-de-tokens-visuais-no-raciocinio-complexo-de-mllm/korshunov.aipt2026-08-11T10:45:00.017168+00:00DSTP mitiga falhas na poda de tokens visuais no raciocínio complexo de MLLM https://korshunov.ai/hi/article/17754-dstp-jttil-mllm-trk-men-drshy-ttokn-pruuning-viphltaaon-ko-km-krtaa-hai/korshunov.aihi2026-08-11T10:45:00.017168+00:00DSTP जटिल MLLM तर्क में दृश्य टोकन प्रूनिंग विफलताओं को कम करता है https://korshunov.ai/ja/article/17754-dstpkafu-za-namllmtui-lun-niokerushi-jue-to-kunhuru-ninkunoshi-bai-wohuan-he/korshunov.aija2026-08-11T10:45:00.017168+00:00DSTPが複雑なMLLM推論における視覚トークンプルーニングの失敗を緩和 https://korshunov.ai/ko/article/17754-dstpga-bogjabhan-mllm-curoneseo-sigagjeog-tokeun-gajicigi-silpaereul/korshunov.aiko2026-08-11T10:45:00.017168+00:00DSTP가 복잡한 MLLM 추론에서 시각적 토큰 가지치기 실패를 완화합니다 https://korshunov.ai/fr/article/17754-dstp-attenue-les-echecs-de-l-elagage-des-tokens-visuels-dans-le-raisonnement-des/korshunov.aifr2026-08-11T10:45:00.017168+00:00DSTP atténue les échecs de l'élagage des tokens visuels dans le raisonnement complexe des MLLM https://korshunov.ai/id/article/17754-dstp-meredam-kegagalan-pemangkasan-token-visual-dalam-penalaran-mllm-yang/korshunov.aiid2026-08-11T10:45:00.017168+00:00DSTP meredam kegagalan pemangkasan token visual dalam penalaran MLLM yang kompleks https://korshunov.ai/ar/article/17754-ykhff-dstp-mn-fshl-tqlym-lrmwz-lbsry-fy-lstdll-lm-qd-lnmdhj-mllm/korshunov.aiar2026-08-11T10:45:00.017168+00:00يخفف DSTP من فشل تقليم الرموز البصرية في الاستدلال المعقد لنماذج MLLM https://korshunov.ai/ko/article/17753-genrecaleun-jaebojeongeul-tonghae-daegyumo-bijeon-eoneo-modeleul-sogyumo-modelro/korshunov.aiko2026-08-11T10:37:10.713058+00:00GenRecal은 재보정을 통해 대규모 비전-언어 모델을 소규모 모델로 증류합니다 https://korshunov.ai/fr/article/17753-genrecal-distille-de-grands-modeles-vision-langage-en-petits-par-recalibration/korshunov.aifr2026-08-11T10:37:10.713058+00:00GenRecal distille de grands modèles vision-langage en petits par recalibration https://korshunov.ai/id/article/17753-genrecal-mendistilasi-model-visi-bahasa-besar-ke-yang-lebih-kecil-melalui/korshunov.aiid2026-08-11T10:37:10.713058+00:00GenRecal mendistilasi model visi-bahasa besar ke yang lebih kecil melalui rekalsibrasi https://korshunov.ai/ar/article/17753-yqwm-genrecal-btqtyr-nmdhj-lrwy-wllg-lkbyr-l-sgyr-br-d-m-yr/korshunov.aiar2026-08-11T10:37:10.713058+00:00يقوم GenRecal بتقطير نماذج الرؤية واللغة الكبيرة إلى صغيرة عبر إعادة معايرة https://korshunov.ai/en/article/17753-genrecal-distills-large-vision-language-models-to-small-ones-via-recalibration/korshunov.aien2026-08-11T10:37:10.713058+00:00GenRecal distills large vision-language models to small ones via recalibration https://korshunov.ai/ru/article/17753-genrecal-distilliruet-krupnye-modeli-zreniia-i-iazyka-v-melkie-posredstvom/korshunov.airu2026-08-11T10:37:10.713058+00:00GenRecal дистиллирует крупные модели зрения и языка в мелкие посредством перекалибровки https://korshunov.ai/es/article/17753-genrecal-destila-grandes-modelos-de-vision-y-lenguaje-en-pequenos-mediante/korshunov.aies2026-08-11T10:37:10.713058+00:00GenRecal destila grandes modelos de visión y lenguaje en pequeños mediante recalibración https://korshunov.ai/zh-hans/article/17753-genrecal-tong-guo-zhong-xin-xiao-zhun-jiang-da-xing-shi-jue-yu-yan-mo-xing-zheng/korshunov.aizh-cn2026-08-11T10:37:10.713058+00:00GenRecal 通过重新校准将大型视觉-语言模型蒸馏为小型模型 https://korshunov.ai/pt-br/article/17753-genrecal-destila-grandes-modelos-de-visao-e-linguagem-em-pequenos-por-meio-de/korshunov.aipt2026-08-11T10:37:10.713058+00:00GenRecal destila grandes modelos de visão e linguagem em pequenos por meio de recalibração https://korshunov.ai/hi/article/17753-genrecal-punh-kailibreshn-ke-maadhym-se-bdde-vijn-laingvej-monddlon-ko-chotte/korshunov.aihi2026-08-11T10:37:10.713058+00:00GenRecal पुनः कैलिब्रेशन के माध्यम से बड़े विज़न-लैंग्वेज मॉडलों को छोटे मॉडलों में डिस्टिल करता है https://korshunov.ai/ja/article/17753-genrecalhazai-jiao-zheng-niyorida-gui-mo-hishiyon-yan-yu-moteruwoxiao-gui-mo-liu/korshunov.aija2026-08-11T10:37:10.713058+00:00GenRecalは再較正により大規模ビジョン・言語モデルを小規模モデルに蒸留する https://korshunov.ai/zh-hans/article/17752-zppo-shi-yong-ti-shi-er-fei-ti-du-lai-gai-jin-xiao-xing-shi-jue-yu-yan-mo-xing/korshunov.aizh-cn2026-08-11T10:29:27.566640+00:00ZPPO 使用提示而非梯度来改进小型视觉语言模型的训练 https://korshunov.ai/en/article/17752-zppo-uses-prompts-instead-of-gradients-to-improve-small-vision-language-model/korshunov.aien2026-08-11T10:29:27.566640+00:00ZPPO uses prompts instead of gradients to improve small vision-language model training https://korshunov.ai/ru/article/17752-zppo-ispolzuet-prompty-vmesto-gradientov-dlia-uluchsheniia-obucheniia-malykh/korshunov.airu2026-08-11T10:29:27.566640+00:00ZPPO использует промпты вместо градиентов для улучшения обучения малых зрительно-языковых моделей https://korshunov.ai/es/article/17752-zppo-utiliza-prompts-en-lugar-de-gradientes-para-mejorar-el-entrenamiento-de/korshunov.aies2026-08-11T10:29:27.566640+00:00ZPPO utiliza prompts en lugar de gradientes para mejorar el entrenamiento de modelos visiolingüísticos pequeños https://korshunov.ai/pt-br/article/17752-zppo-usa-prompts-em-vez-de-gradientes-para-melhorar-o-treinamento-de-modelos-de/korshunov.aipt2026-08-11T10:29:27.566640+00:00ZPPO usa prompts em vez de gradientes para melhorar o treinamento de modelos pequenos de visão e linguagem https://korshunov.ai/hi/article/17752-zppo-chotte-vijn-laingvej-monddl-prshikssnn-ko-behtr-bnaane-ke-lie-greddientt-ke/korshunov.aihi2026-08-11T10:29:27.566640+00:00ZPPO छोटे विज़न-लैंग्वेज मॉडल प्रशिक्षण को बेहतर बनाने के लिए ग्रेडिएंट के बजाय प्रॉम्प्ट का उपयोग करता है https://korshunov.ai/ja/article/17752-zppohagou-pei-nodai-warinihuronhutowoshi-yong-shite-xiao-gui-mo-nahishiyonyan-yu/korshunov.aija2026-08-11T10:29:27.566640+00:00ZPPOは勾配の代わりにプロンプトを使用して、小規模なビジョン言語モデルのトレーニングを改善する https://korshunov.ai/ko/article/17752-zpponeun-jageun-bijeon-eoneo-model-hagseubeul-gaeseonhagi-wihae-geuraedieonteu/korshunov.aiko2026-08-11T10:29:27.566640+00:00ZPPO는 작은 비전-언어 모델 학습을 개선하기 위해 그래디언트 대신 프롬프트를 사용한다 https://korshunov.ai/fr/article/17752-zppo-utilise-des-invites-au-lieu-de-gradients-pour-ameliorer-l-entrainement-de/korshunov.aifr2026-08-11T10:29:27.566640+00:00ZPPO utilise des invites au lieu de gradients pour améliorer l'entraînement de petits modèles vision-langage https://korshunov.ai/id/article/17752-zppo-menggunakan-prompt-alih-alih-gradien-untuk-meningkatkan-pelatihan-model/korshunov.aiid2026-08-11T10:29:27.566640+00:00ZPPO menggunakan prompt alih-alih gradien untuk meningkatkan pelatihan model visi-bahasa kecil https://korshunov.ai/ar/article/17752-ystkhdm-zppo-lmtlbt-bdlan-mn-ltdrjt-lthsyn-tdryb-nmdhj-lrwy-wllg-lsgyr/korshunov.aiar2026-08-11T10:29:27.566640+00:00يستخدم ZPPO المطالبات بدلاً من التدرجات لتحسين تدريب نماذج الرؤية واللغة الصغيرة https://korshunov.ai/hi/article/17751-hide-to-see-ne-vlm-ddisttileshn-ke-lie-trk-puurvakssr-maasking-pesh-kii/korshunov.aihi2026-08-11T10:10:47.279179+00:00Hide to See ने VLM डिस्टिलेशन के लिए तर्क-पूर्वअक्षर मास्किंग पेश की https://korshunov.ai/en/article/17751-hide-to-see-introduces-reasoning-prefix-masking-for-vlm-distillation/korshunov.aien2026-08-11T10:10:47.279179+00:00Hide to See introduces reasoning-prefix masking for VLM distillation https://korshunov.ai/ru/article/17751-hide-to-see-predstavliaet-maskirovanie-prefiksa-rassuzhdeniia-dlia-distilliatsii/korshunov.airu2026-08-11T10:10:47.279179+00:00Hide to See представляет маскирование префикса рассуждения для дистилляции VLM https://korshunov.ai/es/article/17751-hide-to-see-introduce-enmascaramiento-de-prefijo-de-razonamiento-para-vlm/korshunov.aies2026-08-11T10:10:47.279179+00:00Hide to See introduce enmascaramiento de prefijo de razonamiento para destilación VLM https://korshunov.ai/zh-hans/article/17751-hide-to-see-yin-ru-yong-yu-vlm-zheng-liu-de-tui-li-qian-zhui-yan-ma/korshunov.aizh-cn2026-08-11T10:10:47.279179+00:00Hide to See 引入用于 VLM 蒸馏的推理前缀掩码 https://korshunov.ai/pt-br/article/17751-hide-to-see-introduz-mascaramento-de-prefixo-de-raciocinio-para-destilacao-de/korshunov.aipt2026-08-11T10:10:47.279179+00:00Hide to See introduz mascaramento de prefixo de raciocínio para destilação de VLM https://korshunov.ai/ja/article/17751-hide-to-see-ha-vlm-zheng-liu-notamenotui-lun-jie-tou-ci-masukinkuwodao-ru/korshunov.aija2026-08-11T10:10:47.279179+00:00Hide to See は VLM 蒸留のための推論接頭辞マスキングを導入 https://korshunov.ai/ko/article/17751-hide-to-seeneun-vlm-jeungryuyong-curon-jeobdusa-maseukingeul-doibhabnida/korshunov.aiko2026-08-11T10:10:47.279179+00:00Hide to See는 VLM 증류용 추론 접두사 마스킹을 도입합니다 https://korshunov.ai/fr/article/17751-hide-to-see-introduit-le-masquage-du-prefixe-de-raisonnement-pour-la-vlm/korshunov.aifr2026-08-11T10:10:47.279179+00:00Hide to See introduit le masquage du préfixe de raisonnement pour la distillation VLM https://korshunov.ai/id/article/17751-hide-to-see-memperkenalkan-masking-prefix-penalaran-untuk-distilasi-vlm/korshunov.aiid2026-08-11T10:10:47.279179+00:00Hide to See memperkenalkan masking prefix penalaran untuk distilasi VLM https://korshunov.ai/ar/article/17751-hide-to-see-tqdm-qn-bdy-lstdll-ltqtyr-vlm/korshunov.aiar2026-08-11T10:10:47.279179+00:00Hide to See تقدم قناع بادئة الاستدلال لتقطير VLM https://korshunov.ai/en/article/17750-spatialclaw-uses-code-as-action-interface-for-agentic-spatial-reasoning/korshunov.aien2026-08-11T10:02:35.074412+00:00SpatialClaw uses code as action interface for agentic spatial reasoning https://korshunov.ai/ru/article/17750-spatialclaw-ispolzuet-kod-kak-interfeis-deistvii-dlia-agentnogo-rassuzhdeniia/korshunov.airu2026-08-11T10:02:35.074412+00:00SpatialClaw использует код как интерфейс действий для агентного пространственного рассуждения https://korshunov.ai/es/article/17750-spatialclaw-utiliza-codigo-como-interfaz-de-accion-para-el-razonamiento-espacial/korshunov.aies2026-08-11T10:02:35.074412+00:00SpatialClaw utiliza código como interfaz de acción para el razonamiento espacial agéntico https://korshunov.ai/zh-hans/article/17750-spatialclaw-shi-yong-dai-ma-zuo-wei-zhi-neng-ti-kong-jian-tui-li-de-dong-zuo-jie/korshunov.aizh-cn2026-08-11T10:02:35.074412+00:00SpatialClaw 使用代码作为智能体空间推理的动作接口 https://korshunov.ai/pt-br/article/17750-spatialclaw-usa-codigo-como-interface-de-acao-para-raciocinio-espacial-agentico/korshunov.aipt2026-08-11T10:02:35.074412+00:00SpatialClaw usa código como interface de ação para raciocínio espacial agêntico https://korshunov.ai/hi/article/17750-spatialclaw-ejenttik-sthaanik-trk-ke-lie-kriyaa-inttrphes-ke-ruup-men-kodd-kaa/korshunov.aihi2026-08-11T10:02:35.074412+00:00SpatialClaw एजेंटिक स्थानिक तर्क के लिए क्रिया इंटरफ़ेस के रूप में कोड का उपयोग करता है https://korshunov.ai/ja/article/17750-spatialclawhae-shientokong-jian-tui-lun-notamenoakushiyoninta-hue-sutoshiteko/korshunov.aija2026-08-11T10:02:35.074412+00:00SpatialClawはエージェント空間推論のためのアクションインターフェースとしてコードを使用 https://korshunov.ai/ko/article/17750-spatialclawneun-eijeonteu-gonggan-curoneul-wihae-kodeureul-aegsyeon-sayong/korshunov.aiko2026-08-11T10:02:35.074412+00:00SpatialClaw는 에이전트 공간 추론을 위해 코드를 액션 인터페이스로 사용 https://korshunov.ai/fr/article/17750-spatialclaw-utilise-le-code-comme-interface-d-action-pour-le-raisonnement/korshunov.aifr2026-08-11T10:02:35.074412+00:00SpatialClaw utilise le code comme interface d'action pour le raisonnement spatial agentique https://korshunov.ai/id/article/17750-spatialclaw-menggunakan-kode-sebagai-antarmuka-aksi-untuk-penalaran-spasial/korshunov.aiid2026-08-11T10:02:35.074412+00:00SpatialClaw menggunakan kode sebagai antarmuka aksi untuk penalaran spasial agentic https://korshunov.ai/ar/article/17750-ystkhdm-spatialclaw-lkwd-kwjh-jr-llstdll-lmkny-lwkyl/korshunov.aiar2026-08-11T10:02:35.074412+00:00يستخدم SpatialClaw الكود كواجهة إجراء للاستدلال المكاني الوكيل https://korshunov.ai/en/article/17749-axpo-improves-multimodal-agentic-reasoning-by-fixing-the-thinking-acting-gap/korshunov.aien2026-08-11T09:53:58.669949+00:00AXPO improves multimodal agentic reasoning by fixing the Thinking-Acting Gap https://korshunov.ai/ru/article/17749-axpo-uluchshaet-multimodalnoe-agentnoe-rassuzhdenie-ustraniaia-razryv-mezhdu-i/korshunov.airu2026-08-11T09:53:58.669949+00:00AXPO улучшает мультимодальное агентное рассуждение, устраняя разрыв между мышлением и действием https://korshunov.ai/es/article/17749-axpo-mejora-el-razonamiento-agenico-multimodal-corrigiendo-la-brecha-entre-y/korshunov.aies2026-08-11T09:53:58.669949+00:00AXPO mejora el razonamiento agénico multimodal corrigiendo la brecha entre pensar y actuar https://korshunov.ai/zh-hans/article/17749-axpotong-guo-xiu-fu-si-kao-xing-dong-chai-ju-ti-sheng-duo-mo-tai-zhi-neng-ti-tui/korshunov.aizh-cn2026-08-11T09:53:58.669949+00:00AXPO通过修复思考-行动差距提升多模态智能体推理能力 https://korshunov.ai/pt-br/article/17749-axpo-melhora-o-raciocinio-agentico-multimodal-corrigindo-a-lacuna-entre-e-acao/korshunov.aipt2026-08-11T09:53:58.669949+00:00AXPO melhora o raciocínio agêntico multimodal corrigindo a Lacuna entre Pensamento e Ação https://korshunov.ai/hi/article/17749-axpo-ne-thinking-ektting-gaip-ko-tthiik-krke-mlttiimoddl-ejenttik-riijning-men/korshunov.aihi2026-08-11T09:53:58.669949+00:00AXPO ने थिंकिंग-एक्टिंग गैप को ठीक करके मल्टीमोडल एजेंटिक रीजनिंग में सुधार किया https://korshunov.ai/ja/article/17749-axpo-si-kao-toxing-dong-nokiyatsuhuwoxiu-zheng-surukototemaruchimo-tarue-lun/korshunov.aija2026-08-11T09:53:58.669949+00:00AXPO、思考と行動のギャップを修正することでマルチモーダルエージェント推論を改善 https://korshunov.ai/ko/article/17749-axpo-sago-haengdong-gangeugeul-sujeonghayeo-dajungmodal-eijeonteu-curon-gaeseon/korshunov.aiko2026-08-11T09:53:58.669949+00:00AXPO, 사고-행동 간극을 수정하여 다중모달 에이전트 추론 개선 https://korshunov.ai/fr/article/17749-axpo-ameliore-le-raisonnement-agentique-multimodal-en-comblant-l-ecart-entre-et/korshunov.aifr2026-08-11T09:53:58.669949+00:00AXPO améliore le raisonnement agentique multimodal en comblant l'écart entre réflexion et action https://korshunov.ai/id/article/17749-axpo-meningkatkan-penalaran-agentic-multimodal-dengan-memperbaiki-kesenjangan/korshunov.aiid2026-08-11T09:53:58.669949+00:00AXPO meningkatkan penalaran agentic multimodal dengan memperbaiki Kesenjangan Berpikir-Bertindak https://korshunov.ai/ar/article/17749-thswn-axpo-llstdll-lwkyl-mt-dd-lwsyt-br-sd-fjw-ltfkyr-ltnfydh/korshunov.aiar2026-08-11T09:53:58.669949+00:00تحسّن AXPO للاستدلال الوكيل متعدد الوسائط عبر سد فجوة التفكير-التنفيذ https://korshunov.ai/en/article/17744-webai-releases-twil-lm-a-1-7b-and-3b-formal-logic-model-family-for-local/korshunov.aien2026-08-11T09:30:44.989196+00:00webAI releases TwIL-LM, a 1.7B and 3B formal-logic model family for local autoformalization https://korshunov.ai/ru/article/17744-webai-vypuskaet-twil-lm-semeistvo-formalno-logicheskikh-modelei-na-17b-i-3b-dlia/korshunov.airu2026-08-11T09:30:44.989196+00:00webAI выпускает TwIL-LM — семейство формально-логических моделей на 1,7B и 3B параметров для локальной автоформализации https://korshunov.ai/es/article/17744-webai-lanza-twil-lm-una-familia-de-modelos-de-logica-formal-de-1-7b-y-3b-para/korshunov.aies2026-08-11T09:30:44.989196+00:00webAI lanza TwIL-LM, una familia de modelos de lógica formal de 1.7B y 3B para autoformalización local https://korshunov.ai/zh-hans/article/17744-webaifa-bu-twil-lm-mian-xiang-ben-di-zi-dong-xing-shi-hua-de-1-7bhe-3bxing-shi/korshunov.aizh-cn2026-08-11T09:30:44.989196+00:00webAI发布TwIL-LM，面向本地自动形式化的1.7B和3B形式逻辑模型系列 https://korshunov.ai/pt-br/article/17744-webai-lanca-twil-lm-uma-familia-de-modelos-de-logica-formal-de-1-7b-e-3b-para/korshunov.aipt2026-08-11T09:30:44.989196+00:00webAI lança TwIL-LM, uma família de modelos de lógica formal de 1.7B e 3B para autoformalização local https://korshunov.ai/hi/article/17744-webai-ne-sthaaniiy-onttophonrmlaaijeshn-ke-lie-1-7b-aur-3b-phonrml-lonjik-monddl/korshunov.aihi2026-08-11T09:30:44.989196+00:00webAI ने स्थानीय ऑटोफॉर्मलाइजेशन के लिए 1.7B और 3B फॉर्मल-लॉजिक मॉडल परिवार TwIL-LM जारी किया https://korshunov.ai/ja/article/17744-webaikaro-karuzi-dong-xing-shi-hua-xiang-keno1-7boyohi3bnoxing-shi-lun-li-twil/korshunov.aija2026-08-11T09:30:44.989196+00:00webAIがローカル自動形式化向けの1.7Bおよび3Bの形式論理モデルファミリー「TwIL-LM」をリリース https://korshunov.ai/ko/article/17744-webaiga-rokeol-jadong-hyeongsighwareul-wihan-1-7b-mic-3b-gongsig-nonri-model-lm/korshunov.aiko2026-08-11T09:30:44.989196+00:00webAI가 로컬 자동 형식화를 위한 1.7B 및 3B 공식 논리 모델 패밀리 TwIL-LM 출시 https://korshunov.ai/fr/article/17744-webai-publie-twil-lm-une-famille-de-modeles-de-logique-formelle-de-17b-et-3b-l/korshunov.aifr2026-08-11T09:30:44.989196+00:00webAI publie TwIL-LM, une famille de modèles de logique formelle de 1,7B et 3B pour l'auto-formalisation locale https://korshunov.ai/id/article/17744-webai-merilis-twil-lm-keluarga-model-logika-formal-berukuran-17b-dan-3b-untuk/korshunov.aiid2026-08-11T09:30:44.989196+00:00webAI merilis TwIL-LM, keluarga model logika formal berukuran 1,7B dan 3B untuk autoformalisasi lokal https://korshunov.ai/ar/article/17744-wyb-y-y-ttlq-yl-nmdhj-twil-lm-lmntqy-lrsmy-bhjm-1-7b-w3b-lltshkyl-ltlqyy-lmhly/korshunov.aiar2026-08-11T09:30:44.989196+00:00ويب أي آي تطلق عائلة نماذج TwIL-LM المنطقية الرسمية بحجم 1.7B و3B للتشكيل التلقائي المحلي https://korshunov.ai/en/article/17735-meta-releases-muse-glimmer-30b-open-weight-model-anthropic-improves-riemann/korshunov.aien2026-08-11T08:50:44.041940+00:00Meta releases Muse Glimmer 30B open-weight model; Anthropic improves Riemann bound https://korshunov.ai/ru/article/17735-meta-vypustila-model-s-otkrytymi-vesami-muse-glimmer-30b-anthropic-uluchshila/korshunov.airu2026-08-11T08:50:44.041940+00:00Meta выпустила модель с открытыми весами Muse Glimmer 30B; Anthropic улучшила оценку для гипотезы Римана https://korshunov.ai/es/article/17735-meta-lanza-el-modelo-de-peso-abierto-muse-glimmer-30b-anthropic-mejora-la-cota/korshunov.aies2026-08-11T08:50:44.041940+00:00Meta lanza el modelo de peso abierto Muse Glimmer 30B; Anthropic mejora la cota inferior de Riemann https://korshunov.ai/zh-hans/article/17735-metafa-bu-muse-glimmer-30bkai-yuan-quan-zhong-mo-xing-anthropicgai-jin-li-man/korshunov.aizh-cn2026-08-11T08:50:44.041940+00:00Meta发布Muse Glimmer 30B开源权重模型；Anthropic改进黎曼边界 https://korshunov.ai/pt-br/article/17735-meta-lanca-modelo-de-pesos-abertos-muse-glimmer-30b-anthropic-melhora-limite-de/korshunov.aipt2026-08-11T08:50:44.041940+00:00Meta lança modelo de pesos abertos Muse Glimmer 30B; Anthropic melhora limite de Riemann https://korshunov.ai/hi/article/17735-mettaa-ne-muse-glimmer-30b-opn-vett-monddl-jaarii-kiyaa-anthropic-ne-riemann-men/korshunov.aihi2026-08-11T08:50:44.041940+00:00मेटा ने Muse Glimmer 30B ओपन-वेट मॉडल जारी किया; Anthropic ने Riemann बॉउंड में सुधार किया https://korshunov.ai/ja/article/17735-metakamuse-glimmer-30bnoo-hunueitomoteruworiri-su-anthropickari-mannojing-jie/korshunov.aija2026-08-11T08:50:44.041940+00:00MetaがMuse Glimmer 30Bのオープンウェイトモデルをリリース、Anthropicがリーマンの境界を改善 https://korshunov.ai/ko/article/17735-meta-muse-glimmer-30b-opeun-gajungci-model-culsi-anthropic-riman-gyeonggye/korshunov.aiko2026-08-11T08:50:44.041940+00:00Meta, Muse Glimmer 30B 오픈 가중치 모델 출시; Anthropic, 리만 경계 개선 https://korshunov.ai/fr/article/17735-meta-publie-le-modele-a-poids-ouverts-muse-glimmer-30b-anthropic-ameliore-la-de/korshunov.aifr2026-08-11T08:50:44.041940+00:00Meta publie le modèle à poids ouverts Muse Glimmer 30B ; Anthropic améliore la borne de Riemann https://korshunov.ai/id/article/17735-meta-merilis-model-open-weight-muse-glimmer-30b-anthropic-meningkatkan-batas/korshunov.aiid2026-08-11T08:50:44.041940+00:00Meta merilis model open-weight Muse Glimmer 30B; Anthropic meningkatkan batas Riemann https://korshunov.ai/ar/article/17735-tlq-myt-lmwdyl-muse-glimmer-30b-dhw-lwzn-lmftwh-nthrwbyk-thswn-lhd-ldn-l-riemann/korshunov.aiar2026-08-11T08:50:44.041940+00:00إطلاق ميتا لموديل Muse Glimmer 30B ذو الأوزان المفتوحة؛ أنثروبيك تحسّن الحد الأدنى لـ Riemann https://korshunov.ai/en/article/17734-meta-releases-muse-glimmer-a-30b-open-weight-model-for-local-agents/korshunov.aien2026-08-11T08:24:07.578417+00:00Meta releases Muse Glimmer, a 30B open-weight model for local agents https://korshunov.ai/ru/article/17734-meta-vypustila-muse-glimmer-30b-model-s-otkrytym-vesom-dlia-lokalnykh-agentov/korshunov.airu2026-08-11T08:24:07.578417+00:00Meta выпустила Muse Glimmer, 30B модель с открытым весом для локальных агентов https://korshunov.ai/es/article/17734-meta-lanza-muse-glimmer-un-modelo-de-peso-abierto-de-30b-para-agentes-locales/korshunov.aies2026-08-11T08:24:07.578417+00:00Meta lanza Muse Glimmer, un modelo de peso abierto de 30B para agentes locales https://korshunov.ai/zh-hans/article/17734-metafa-bu-muse-glimmer-yi-kuan-mian-xiang-ben-di-dai-li-de-30bkai-yuan-quan-mo/korshunov.aizh-cn2026-08-11T08:24:07.578417+00:00Meta发布Muse Glimmer，一款面向本地代理的30B开源权重模型 https://korshunov.ai/pt-br/article/17734-meta-lanca-muse-glimmer-um-modelo-de-peso-aberto-de-30b-para-agentes-locais/korshunov.aipt2026-08-11T08:24:07.578417+00:00Meta lança Muse Glimmer, um modelo de peso aberto de 30B para agentes locais https://korshunov.ai/hi/article/17734-meta-ne-sthaaniiy-ejentts-ke-lie-30b-opn-vett-monddl-muse-glimmer-jaarii-kiyaa/korshunov.aihi2026-08-11T08:24:07.578417+00:00Meta ने स्थानीय एजेंट्स के लिए 30B ओपन-वेट मॉडल Muse Glimmer जारी किया https://korshunov.ai/ja/article/17734-metakaro-karue-shientoxiang-keno30bo-hunueitomoteru-muse-glimmer-woriri-su/korshunov.aija2026-08-11T08:24:07.578417+00:00Metaがローカルエージェント向けの30Bオープンウェイトモデル「Muse Glimmer」をリリース https://korshunov.ai/ko/article/17734-meta-rokeol-eijeonteureul-wihan-30b-opeun-gajungci-model-muse-glimmer-culsi/korshunov.aiko2026-08-11T08:24:07.578417+00:00Meta, 로컬 에이전트를 위한 30B 오픈 가중치 모델 Muse Glimmer 출시 https://korshunov.ai/fr/article/17734-meta-lance-muse-glimmer-un-modele-de-poids-ouvert-de-30b-pour-agents-locaux/korshunov.aifr2026-08-11T08:24:07.578417+00:00Meta lance Muse Glimmer, un modèle de poids ouvert de 30B pour agents locaux https://korshunov.ai/id/article/17734-meta-merilis-muse-glimmer-model-bobot-terbuka-30b-untuk-agen-lokal/korshunov.aiid2026-08-11T08:24:07.578417+00:00Meta merilis Muse Glimmer, model bobot terbuka 30B untuk agen lokal https://korshunov.ai/ar/article/17734-myt-ttlq-muse-glimmer-nmwdhj-bwzn-mftwh-bhjm-30b-llwkl-lmhlyyn/korshunov.aiar2026-08-11T08:24:07.578417+00:00مايتا تطلق Muse Glimmer، نموذج بوزن مفتوح بحجم 30B للوكلاء المحليين https://korshunov.ai/es/article/17727-macaron-v1-presenta-una-familia-de-modelos-agente-abierta-con-mixture-of-lora/korshunov.aies2026-08-11T08:05:52.262695+00:00Macaron-V1 presenta una familia de modelos-agente abierta con Mixture-of-LoRA para aprendizaje continuo https://korshunov.ai/en/article/17727-macaron-v1-introduces-open-agent-model-family-with-mixture-of-lora-for-continual/korshunov.aien2026-08-11T08:05:52.262695+00:00Macaron-V1 introduces open agent-model family with Mixture-of-LoRA for continual learning https://korshunov.ai/ru/article/17727-macaron-v1-predstavliaet-otkrytoe-semeistvo-agentnykh-modelei-s-mixture-of-lora/korshunov.airu2026-08-11T08:05:52.262695+00:00Macaron-V1 представляет открытое семейство агентных моделей с Mixture-of-LoRA для непрерывного обучения https://korshunov.ai/zh-hans/article/17727-macaron-v1-yin-ru-ji-yu-mixture-of-lora-de-chi-xu-xue-xi-kai-fang-zhi-neng-ti-mo/korshunov.aizh-cn2026-08-11T08:05:52.262695+00:00Macaron-V1 引入基于 Mixture-of-LoRA 的持续学习开放智能体模型家族 https://korshunov.ai/pt-br/article/17727-macaron-v1-apresenta-familia-de-modelos-de-agente-aberta-com-mixture-of-lora/korshunov.aipt2026-08-11T08:05:52.262695+00:00Macaron-V1 apresenta família de modelos de agente aberta com Mixture-of-LoRA para aprendizado contínuo https://korshunov.ai/hi/article/17727-macaron-v1-ne-mixture-of-lora-ke-saath-nirntr-siikhne-ke-lie-khulaa-ejentt-pesh/korshunov.aihi2026-08-11T08:05:52.262695+00:00Macaron-V1 ने Mixture-of-LoRA के साथ निरंतर सीखने के लिए खुला एजेंट-मॉडल परिवार पेश किया https://korshunov.ai/ja/article/17727-macaron-v1kaji-sok-xue-xi-notamenomixture-of-lorawoyong-itao-hunnae-wodao-ru/korshunov.aija2026-08-11T08:05:52.262695+00:00Macaron-V1が継続学習のためのMixture-of-LoRAを用いたオープンなエージェントモデルファミリーを導入 https://korshunov.ai/ko/article/17727-macaron-v1i-mixture-of-lorareul-hwalyonghan-jisog-hagseubyong-opeun-eijeonteu/korshunov.aiko2026-08-11T08:05:52.262695+00:00Macaron-V1이 Mixture-of-LoRA를 활용한 지속 학습용 오픈 에이전트 모델 패밀리 출시 https://korshunov.ai/fr/article/17727-macaron-v1-presente-une-famille-de-modeles-d-agent-ouverte-avec-mixture-of-lora/korshunov.aifr2026-08-11T08:05:52.262695+00:00Macaron-V1 présente une famille de modèles d'agent ouverte avec Mixture-of-LoRA pour l'apprentissage continu https://korshunov.ai/id/article/17727-macaron-v1-memperkenalkan-keluarga-agen-model-terbuka-dengan-mixture-of-lora/korshunov.aiid2026-08-11T08:05:52.262695+00:00Macaron-V1 memperkenalkan keluarga agen-model terbuka dengan Mixture-of-LoRA untuk pembelajaran berkelanjutan https://korshunov.ai/ar/article/17727-macaron-v1-yutlq-yl-nmdhj-wkl-mftwh-lmsdr-bstkhdm-mixture-of-lora-llt-lm-lmstmr/korshunov.aiar2026-08-11T08:05:52.262695+00:00Macaron-V1 يُطلق عائلة نماذج وكلاء مفتوحة المصدر باستخدام Mixture-of-LoRA للتعلم المستمر https://korshunov.ai/en/article/17720-researchers-exploit-cross-session-encryption-reuse-to-steal-reasoning-traces-llm/korshunov.aien2026-08-11T07:44:16.124096+00:00Researchers exploit cross-session encryption reuse to steal reasoning traces from LLM APIs https://korshunov.ai/ja/article/17720-yan-jiu-zhe-kakurosusetsushiyonan-hao-hua-nozai-li-yong-woe-yong-shi-llm-api-lun/korshunov.aija2026-08-11T07:44:16.124096+00:00研究者がクロスセッション暗号化の再利用を悪用し、LLM API から推論トレースを窃取 https://korshunov.ai/ru/article/17720-issledovateli-ispolzuiut-povtornoe-primenenie-skvoznoi-sessiinoi-shifrovaniia-iz/korshunov.airu2026-08-11T07:44:16.124096+00:00Исследователи используют повторное применение сквозной сессийной шифрования для кражи цепочек рассуждений из API LLM https://korshunov.ai/ar/article/17720-bhthwn-ystglwn-d-stkhdm-ltshfyr-br-ljlst-lshb-msrt-lstdll-mn-wjht-brmj-ttbyqt/korshunov.aiar2026-08-11T07:44:16.124096+00:00باحثون يستغلون إعادة استخدام التشفير عبر الجلسات لسحب مسارات الاستدلال من واجهات برمجة تطبيقات نماذج اللغة الكبيرة https://korshunov.ai/es/article/17720-investigadores-explotan-la-reutilizacion-de-cifrado-entre-sesiones-para-robar-de/korshunov.aies2026-08-11T07:44:16.124096+00:00Investigadores explotan la reutilización de cifrado entre sesiones para robar trazas de razonamiento de APIs de LLM https://korshunov.ai/zh-hans/article/17720-yan-jiu-ren-yuan-li-yong-kua-hui-hua-jia-mi-zhong-yong-cong-da-yu-yan-mo-xing-qu/korshunov.aizh-cn2026-08-11T07:44:16.124096+00:00研究人员利用跨会话加密重用从大语言模型API中窃取推理轨迹 https://korshunov.ai/pt-br/article/17720-pesquisadores-exploram-reutilizacao-de-criptografia-entre-sessoes-para-roubar-de/korshunov.aipt2026-08-11T07:44:16.124096+00:00Pesquisadores exploram reutilização de criptografia entre sessões para roubar rastros de raciocínio de APIs de LLM https://korshunov.ai/hi/article/17720-shodhkrtaa-krons-seshn-enkripshn-pun-upyog-kaa-laabh-utthaakr-llm-api-se-trk/korshunov.aihi2026-08-11T07:44:16.124096+00:00शोधकर्ता क्रॉस-सेशन एन्क्रिप्शन पुन: उपयोग का लाभ उठाकर LLM API से तर्क ट्रेस चुराते हैं https://korshunov.ai/ko/article/17720-yeongujadeuli-sesyeon-gan-amhohwa-jaesayongeul-agyonghayeo-llm-apieseo-curon/korshunov.aiko2026-08-11T07:44:16.124096+00:00연구자들이 세션 간 암호화 재사용을 악용하여 LLM API에서 추론 추적을 탈취 https://korshunov.ai/fr/article/17720-des-chercheurs-exploitent-la-reutilisation-du-chiffrement-inter-sessions-pour-de/korshunov.aifr2026-08-11T07:44:16.124096+00:00Des chercheurs exploitent la réutilisation du chiffrement inter-sessions pour voler les traces de raisonnement des API de LLM https://korshunov.ai/id/article/17720-peneliti-mengeksploitasi-penggunaan-ulang-enkripsi-lintas-sesi-untuk-mencuri-api/korshunov.aiid2026-08-11T07:44:16.124096+00:00Peneliti mengeksploitasi penggunaan ulang enkripsi lintas sesi untuk mencuri jejak penalaran dari API LLM https://korshunov.ai/en/article/17707-open-ended-optimization-lets-gpt-5-5-compose-its-own-improvement-process/korshunov.aien2026-08-11T07:16:41.279783+00:00Open-Ended Optimization lets GPT-5.5 compose its own improvement process https://korshunov.ai/ru/article/17707-otkrytaia-optimizatsiia-pozvoliaet-gpt-5-5-sozdavat-sobstvennyi-protsess/korshunov.airu2026-08-11T07:16:41.279783+00:00Открытая оптимизация позволяет GPT-5.5 создавать собственный процесс улучшения https://korshunov.ai/es/article/17707-la-optimizacion-de-proposito-abierto-permite-a-gpt-5-5-componer-su-propio-de/korshunov.aies2026-08-11T07:16:41.279783+00:00La optimización de propósito abierto permite a GPT-5.5 componer su propio proceso de mejora https://korshunov.ai/zh-hans/article/17707-kai-fang-shi-you-hua-shi-gpt-5-5neng-gou-zi-xing-gou-jian-gai-jin-liu-cheng/korshunov.aizh-cn2026-08-11T07:16:41.279783+00:00开放式优化使GPT-5.5能够自行构建改进流程 https://korshunov.ai/pt-br/article/17707-otimizacao-de-finalidade-aberta-permite-que-o-gpt-5-5-componha-seu-proprio-de/korshunov.aipt2026-08-11T07:16:41.279783+00:00Otimização de Finalidade Aberta permite que o GPT-5.5 componha seu próprio processo de melhoria https://korshunov.ai/hi/article/17707-opn-enddedd-onpttimaaijeshn-gpt-5-5-ko-apnii-sudhaar-prkriyaa-svyn-snyojit-krne/korshunov.aihi2026-08-11T07:16:41.279783+00:00ओपन-एंडेड ऑप्टिमाइजेशन GPT-5.5 को अपनी सुधार प्रक्रिया स्वयं संयोजित करने देता है https://korshunov.ai/ja/article/17707-o-hunentozui-shi-hua-niyorigpt-5-5kazi-shen-nogai-shan-hurosesuwogou-zhu/korshunov.aija2026-08-11T07:16:41.279783+00:00オープンエンド最適化によりGPT-5.5が自身の改善プロセスを構築 https://korshunov.ai/ko/article/17707-opeun-endideu-coejeoghwaga-gpt-5-5ege-jace-gaeseon-gwajeong-jagseongeul-hada/korshunov.aiko2026-08-11T07:16:41.279783+00:00오픈 엔디드 최적화가 GPT-5.5에게 자체 개선 과정 작성을 가능하게 하다 https://korshunov.ai/fr/article/17707-l-optimisation-a-finalite-ouverte-permet-a-gpt-5-5-de-composer-son-propre-d/korshunov.aifr2026-08-11T07:16:41.279783+00:00L'optimisation à finalité ouverte permet à GPT-5.5 de composer son propre processus d'amélioration https://korshunov.ai/id/article/17707-optimisasi-terbuka-memungkinkan-gpt-5-5-menyusun-proses-perbaikannya-sendiri/korshunov.aiid2026-08-11T07:16:41.279783+00:00Optimisasi Terbuka Memungkinkan GPT-5.5 menyusun proses perbaikannya sendiri https://korshunov.ai/ar/article/17707-lthsyn-lmftwh-lnhy-ytyh-l-gpt-5-5-tlyf-mly-thsynh-lkhs/korshunov.aiar2026-08-11T07:16:41.279783+00:00التحسين المفتوح النهاية يتيح لـ GPT-5.5 تأليف عملية تحسينه الخاصة https://korshunov.ai/zh-hans/article/17675-amie-video-zai-shi-shi-yi-liao-zi-xun-zhong-da-dao-zhuan-jia-ji-biao-xian/korshunov.aizh-cn2026-08-11T06:26:44.727065+00:00AMIE (Video) 在实时医疗咨询中达到专家级表现 https://korshunov.ai/pt-br/article/17675-amie-video-atinge-desempenho-de-nivel-especialista-em-consultas-medicas-em-tempo/korshunov.aipt2026-08-11T06:26:44.727065+00:00AMIE (Video) atinge desempenho de nível especialista em consultas médicas em tempo real https://korshunov.ai/hi/article/17675-amie-video-vaastvik-smy-cikitsaa-praamrshon-men-vishessjny-str-kaa-prdrshn-krtaa/korshunov.aihi2026-08-11T06:26:44.727065+00:00AMIE (Video) वास्तविक समय चिकित्सा परामर्शों में विशेषज्ञ-स्तर का प्रदर्शन प्राप्त करता है https://korshunov.ai/en/article/17675-amie-video-achieves-expert-level-performance-in-real-time-medical-consultations/korshunov.aien2026-08-11T06:26:44.727065+00:00AMIE (Video) achieves expert-level performance in real-time medical consultations https://korshunov.ai/ru/article/17675-amie-video-demonstriruet-ekspertnyi-uroven-v-realnom-vremeni-pri-meditsinskikh/korshunov.airu2026-08-11T06:26:44.727065+00:00AMIE (Video) демонстрирует экспертный уровень в реальном времени при медицинских консультациях https://korshunov.ai/es/article/17675-amie-video-alcanza-un-rendimiento-de-nivel-experto-en-consultas-medicas-en-real/korshunov.aies2026-08-11T06:26:44.727065+00:00AMIE (Video) alcanza un rendimiento de nivel experto en consultas médicas en tiempo real https://korshunov.ai/ja/article/17675-amie-video-kariarutaimunoyi-liao-xiang-tan-teekisuha-toreherunohahuo-mansuwoda/korshunov.aija2026-08-11T06:26:44.727065+00:00AMIE (Video) がリアルタイムの医療相談でエキスパートレベルのパフォーマンスを達成 https://korshunov.ai/ko/article/17675-amie-video-ga-silsigan-yiryo-sangdameseo-jeonmunga-sujunyi-seongneung-dalseong/korshunov.aiko2026-08-11T06:26:44.727065+00:00AMIE (Video)가 실시간 의료 상담에서 전문가 수준의 성능 달성 https://korshunov.ai/fr/article/17675-amie-video-atteint-des-performances-de-niveau-expert-dans-les-consultations-en/korshunov.aifr2026-08-11T06:26:44.727065+00:00AMIE (Video) atteint des performances de niveau expert dans les consultations médicales en temps réel https://korshunov.ai/id/article/17675-amie-video-mencapai-kinerja-tingkat-ahli-dalam-konsultasi-medis-real-time/korshunov.aiid2026-08-11T06:26:44.727065+00:00AMIE (Video) mencapai kinerja tingkat ahli dalam konsultasi medis real-time https://korshunov.ai/ar/article/17675-amie-video-yhqq-dan-l-mstw-lkhbr-fy-lstshrt-ltby-fy-lwqt-lf-ly/korshunov.aiar2026-08-11T06:26:44.727065+00:00AMIE (Video) يحقق أداءً على مستوى الخبراء في الاستشارات الطبية في الوقت الفعلي https://korshunov.ai/en/article/17674-attackers-steal-reasoning-traces-from-proprietary-llm-apis-via-encrypted-block/korshunov.aien2026-08-11T06:19:04.193031+00:00Attackers steal reasoning traces from proprietary LLM APIs via encrypted block vulnerability https://korshunov.ai/ru/article/17674-zloumyshlenniki-kradut-tsepochki-rassuzhdenii-iz-proprietarnykh-llm-api-cherez-v/korshunov.airu2026-08-11T06:19:04.193031+00:00Злоумышленники крадут цепочки рассуждений из проприетарных LLM API через уязвимость в шифрованных блоках https://korshunov.ai/es/article/17674-los-atacantes-roban-trazas-de-razonamiento-de-apis-de-llm-propietarios-mediante/korshunov.aies2026-08-11T06:19:04.193031+00:00Los atacantes roban trazas de razonamiento de APIs de LLM propietarios mediante una vulnerabilidad en bloques cifrados https://korshunov.ai/zh-hans/article/17674-gong-ji-zhe-tong-guo-jia-mi-kuai-lou-dong-cong-zhuan-you-da-yu-yan-mo-xing-qie/korshunov.aizh-cn2026-08-11T06:19:04.193031+00:00攻击者通过加密块漏洞从专有大语言模型API中窃取推理轨迹 https://korshunov.ai/pt-br/article/17674-atacantes-roubam-rastros-de-raciocinio-de-apis-de-llm-proprietarias-por-meio-de/korshunov.aipt2026-08-11T06:19:04.193031+00:00Atacantes roubam rastros de raciocínio de APIs de LLM proprietárias por meio de vulnerabilidade em blocos criptografados https://korshunov.ai/hi/article/17674-aakraamk-enkripttedd-blonk-kmjorii-ke-maadhym-se-propraaittrii-llm-apis-se-trk/korshunov.aihi2026-08-11T06:19:04.193031+00:00आक्रामक एन्क्रिप्टेड ब्लॉक कमजोरी के माध्यम से प्रोप्राइटरी LLM APIs से तर्क ट्रेस चुराते हैं https://korshunov.ai/ja/article/17674-gong-ji-zhe-kaan-hao-hua-hurotsukunocui-ruo-xing-wotong-shitehurohuraietarinallm/korshunov.aija2026-08-11T06:19:04.193031+00:00攻撃者が暗号化ブロックの脆弱性を通じてプロプライエタリなLLM APIから推論トレースを窃取 https://korshunov.ai/ko/article/17674-gonggyeogjaga-amhohwadoen-beulrog-cwiyagjeomeul-tonghae-dogjeom-llm-apieseo/korshunov.aiko2026-08-11T06:19:04.193031+00:00공격자가 암호화된 블록 취약점을 통해 독점 LLM API에서 추론 추적을 탈취 https://korshunov.ai/fr/article/17674-des-attaquants-volent-des-traces-de-raisonnement-depuis-les-api-de-llm-via-une/korshunov.aifr2026-08-11T06:19:04.193031+00:00Des attaquants volent des traces de raisonnement depuis les API de LLM propriétaires via une vulnérabilité de blocs chiffrés https://korshunov.ai/id/article/17674-penyerang-mencuri-jejak-penalaran-dari-api-llm-tertutup-melalui-kerentanan-blok/korshunov.aiid2026-08-11T06:19:04.193031+00:00Penyerang mencuri jejak penalaran dari API LLM tertutup melalui kerentanan blok terenkripsi https://korshunov.ai/ar/article/17674-mhjmwn-ysrqwn-msrt-lstdll-mn-wjht-brmj-ttbyqt-lnmdhj-llgwy-lkbyr-lkhs-br-thgr-fy/korshunov.aiar2026-08-11T06:19:04.193031+00:00مهاجمون يسرقون مسارات الاستدلال من واجهات برمجة تطبيقات النماذج اللغوية الكبيرة الخاصة عبر ثغرة في الكتلة المشفرة