media MarkTechPost · 1h ago · releases

Anthropic, OpenAI, and Google DeepMind release four frontier models in September

from English

Anthropic, OpenAI, and Google DeepMind launched four frontier-class models within a 30-day period in September 2026: Claude Fable 5.1, GPT-6 Astra, GPT-6.1 Sol, and Gemini 4 Argon.

Claude Fable 5.1 released on September 1 with $10/$50 pricing and 1M context window.
GPT-6 Astra followed on September 3 as OpenAI's top model for software engineering and computer use tasks.
GPT-6.1 Sol launched on September 29 at one-fifth the cost of Astra, matching it on DeepSWE v1.1 benchmarks.
Gemini 4 Argon announced on September 30 offers a 1M output token cap and leads in knowledge-work benchmarks like Vals Index.
OpenAI cancelled GPT-6.1 Astra after internal scope and authorization tests failed.

The comparison highlights that while benchmark scores overlap significantly, pricing structures and access rules differ substantially, with Sol and Argon offering lower-cost alternatives for high-volume tasks.

Importance 3/3 MarkTechPost OpenAI Anthropic Google DeepMind API & product launches Benchmark results

Benchmarks

Benchmark	Model	Score
ARC-AGI 2	GPT-6 Astra	95%
ARC-AGI 2	Claude Fable 5.1	90%

Read original

media Don't Worry About the Vase · 9d ago · 36 views

Anthropic releases Claude Opus 5.5 with Fable 5.1-level performance at lower cost

Anthropic has introduced Claude Opus 5.5, the first model in its new Claude 5.5 family, which delivers performance comparable to Claude Fable 5.1 on most tasks while costing 40% less to run than Opus 5.

media TLDR AI · 12d ago · 45 views

OpenAI releases GPT-6 Sol and Luna; Anthropic launches Claude Opus 5.5

OpenAI introduced GPT-6 Sol and Luna as faster, more affordable counterparts to GPT-6 Astra, bringing advances in coding, factuality, computer use, and professional tasks to lower-cost models.

media Latent Space · 12d ago · 52 views

Anthropic launches Claude Opus 5.5 with improved writing and lower costs

Anthropic has released Claude Opus 5.5, the first model in its new Claude 5.5 family, positioning it as a more efficient alternative to previous generations. The model is designed to perform at the level of Claude Fable 5.1 for most tasks while costing 40% less to run than Opus 5.

media TLDR AI · 52d ago · 102 views

OpenAI previews GPT-5.6 Sol Ultrafast mode; Google releases Gemini 3.7 Flash

OpenAI has previewed Ultrafast, a GPT-5.6 Sol mode capable of generating up to 750 output tokens per second and operating at up to 14x standard processing speed. The approach aims to deliver real-time performance without switching to a smaller model.

arxiv arXiv cs.AI · 53d ago · 56 views

VITA clinical RAG matches or outperforms frontier LLMs on HealthBench

A purpose-built retrieval-augmented generation system named VITA, designed for low- and middle-income settings, was evaluated against general-purpose large language models on the HealthBench benchmark. The study demonstrates that corpus-specific design can maintain competitive performance even as newer frontier models are released.