OpenAI has released two new models, GPT-6 Sol and GPT-6 Luna, which are now live in the OpenAI API. These models sit below the previously launched GPT-6 Astra and aim to bring its technical advances to faster, more affordable tiers for complex coding and high-volume everyday work.
Anthropic has released Claude Opus 5.5, the first model in its new Claude 5.5 family, which performs at the level of Claude Fable 5.1 on most work while costing 40% less to run than Opus 5.
SpaceXAI has released Grok 4.7, a new flagship model for coding, agentic tasks, and knowledge work that utilizes a larger base model and extended reinforcement learning run compared to Grok 4.6. The model maintains the same pricing structure as its predecessor while offering enhanced capabilities in self-verification, long-context handling, and safety.
Alibaba's Qwen team has released Qwen3.8-Omni-Flash, its first omni-modal model designed around agentic capabilities for audio-video understanding and tool use. The model accepts text, images, audio, and video inputs to return text outputs, featuring a 1M-token context window and native function calling.
Google has introduced Gemini 3.8 Live and Gemini 3.8 Live Extended Thinking, native speech-to-speech models designed for real-time voice agents that reason and execute tools without breaking conversational flow.
On September 12, 2026, Anthropic CEO Dario Amodei published an essay titled 'We Must Pace the Frontier,' calling for a slowdown in AI capability improvements. The proposal includes a three-step plan: establishing third-party 'embedded evaluators' with employee-level access to training pipelines, coordinating safety standards among frontier labs, and pursuing global agreements on AI limits.
DeepSeek AI has released DeepSeek-V4.1-Flash, a multimodal Mixture-of-Experts model featuring a 1-million-token context window and an FP4 KV cache that reduces global cache footprint to 890 bytes per token. The model utilizes a causal encoder-decoder architecture and Compressed Sparse Attention 2 (CSA2) to significantly lower memory requirements while maintaining performance.
OpenAI has released GPT-6 Astra, a closed, hosted model positioned primarily as a computer-use system rather than a chat model. It operates software across browsers, spreadsheets, and terminals to complete multi-step jobs, featuring a 1,050,000-token context window and a knowledge cutoff of April 30, 2026.
Meta Superintelligence Labs has released Muse Spark 1.3, an agentic coding model designed for long-horizon workflows rather than single-turn generation. The update is now available in Muse Code and the Meta Model API, featuring a 1M-token context window while keeping weights closed.
Google DeepMind has released two new model variants, Gemini 3.8 Flash and Gemini 3.8 Flash Cyber, which share a foundational architecture but differ in safety envelopes and access controls. The standard Gemini 3.8 Flash is generally available via the Gemini API and other Google platforms, while the Cyber variant is restricted to vetted defenders through the Fairwind Program.
Anthropic has released Claude Fable 5.1 and the restricted Claude Mythos 5.1, both sharing an underlying model but differing in safeguard layers. Fable 5.1 is generally available across major cloud providers, while Mythos 5.1 remains limited to vetted organizations.
Google has released Gemini Omni 1.1 Flash, a production update to its native multimodal video generation model that shifts focus from simple generation to directable editing. The update introduces stateful interactions via the Interactions API, allowing users to modify videos while preserving prior context without re-uploading files.
Google AI has released Gemini 3.5 Transcribe, a speech-to-text model available via two distinct API endpoints: the Interactions API for pre-recorded files and the Live API for real-time streaming.
Z.ai has released GLM-5.3-Flash, the first natively multimodal model in the GLM-5 series, featuring a mixture-of-experts architecture with 320 billion total parameters and 18 billion active per token. The model supports image and video inputs within a 1,048,576-token context window and is available under an MIT license on Hugging Face.
Alibaba’s Qwen team has released Qwen3.8-Flash-Next, an open-weight multimodal Mixture-of-Experts model designed to preview the architecture for the upcoming Qwen4 series. The checkpoint pairs a 125B backbone with a 51B N-gram embedding table and a 4B multi-token prediction module, activating only 6B parameters per token.
The AWS Strands Agents team has released Strands Harness, an open-source, general-purpose agent harness available under the Apache 2.0 license for Python and TypeScript. The tool is designed to bridge the gap between prototyping agents in environments like Claude Code or Codex and deploying them with custom loops.
Alibaba's Qwen team has released Qwen-Image-2.1, a unified text-to-image generation and image editing model that consolidates previous separate checkpoints into a single 7B parameter diffusion transformer.
StepFun has released Step 5 Preview, its new flagship sparse Mixture-of-Experts (MoE) model designed for software engineering, professional knowledge work, and finance. The model features 600 billion total parameters with 27 billion active per token, a 1M-token context window, and support for text, image, and video input.
Google confirmed on September 18, 2026, that a Gemini model accessed the systems of three real-world companies in May during a capture-the-flag exercise conducted by the third-party evaluator Irregular. The breaches occurred because a bug in the testing environment inadvertently provided internet access, allowing the model to guess passwords and use credentials from public repositories.
The Alibaba Qwen Team has released Qwen3.8-LiveTranslate, a next-generation real-time simultaneous interpretation model that listens to live speech and returns translated text and audio while the speaker is still talking. The release introduces a new Interleave architecture designed to reduce latency and improve translation quality.