OpenAI introduced GPT-6 Sol and GPT-6 Luna on September 22, 2026, two models built on methods behind GPT-6 Astra that cut API prices by 50% versus their GPT-5.6 counterparts. GPT-6 Sol is priced at $2…
ZeroDrift Inc. launched Anchor 3.0, a family of three small language models that check AI-agent-generated messages against financial regulations and corporate policies in real time, available through …
An arXiv paper posted on 15 September 2026 presents an exposition of a proof of the Erdős–Sós Conjecture, which states that every graph with average degree greater than t-2 contains every tree on t≥2 …
OpenAI removed the option for zero reasoning effort in GPT-6 Astra, mandating that every API call consume reasoning tokens billed at the premium output rate of $50 per million tokens, while Anthropic'…
A developer reports that recent frontier model upgrades from OpenAI and Anthropic, including the forthcoming GPT-6 Astra and Claude Opus 5.5, tend to overengineer simple tasks, turning basic framework…
Parallel cut its research time and cost by 50% after moving a production research-agent workload to OpenAI's GPT-6 Astra, according to OpenAI. The same multi-step retrieval-and-synthesis pipeline now …
A developer benchmarked GPT-6 Sol, GPT-6 Astra, and Claude Opus 5.5 on three web apps built from identical English briefs with no prototype or design system, finding that Opus 5.5 consistently priorit…
A new benchmark called FrontierMath Erdős (FME), introduced in a paper submitted to arXiv on 6 September 2026, evaluates AI systems on 68 Erdős problems that remained open as of August 2026, requiring…
Higgsfield AI posted samurai-game demos on September 22nd labeled Claude Opus 5.5 and GPT-6 Astra, using the two rival models to showcase its own game-building and asset-generation tools rather than t…
OpenAI released GPT-6 Sol, a cost-efficient high-end model in its GPT-6 series positioned below the flagship GPT-6 Astra and above the fast GPT-6 Luna tier, priced at $2-4 per million input tokens and…
OpenAI's GPT-6 Astra achieved a 95% success rate on a physical zero-shot pick-and-place task, completing 19 of 20 attempts with dual bimanual I2RT YAM robot arms in evaluations conducted by third-part…
OpenAI released GPT-6 Astra on 2026-09-04, according to a model listing on Amazon Bedrock under the identifier us.openai.gpt-6-astra. The model accepts text and image input with a 1,050k-token context…
OpenAI released GPT-6 Astra globally on 2026-09-04, according to a provider announcement listed on Amazon Bedrock under the identifier global.openai.gpt-6-astra. The model accepts text and image input…
TypeSafe's new System One model, Jev, was adopted by the open-source Matrix agent platform MindRoom within two days of release, where it now powers three decisions including adaptive agent participati…
In a blind evaluation on the How I AI podcast, host Claire Vo tested Claude Opus 5.5, GPT-6 Astra, GPT-6 Sol, and other newly launched models across emails, PRDs, frontend prototypes, backend work, lo…
Anthropic PBC released Claude Opus 5.5 and cut its price 20%, to $4 per million input tokens and $20 per million output, while OpenAI Group PBC released two cheaper GPT-6 models, Sol and Luna, minutes…
OpenAI launched GPT-6 Sol and Luna on Tuesday, two lighter-weight models in its GPT-6 family priced at about half the token cost of their GPT-5.6 counterparts. OpenAI said the models were trained usin…
LinearSolveBench launched as a benchmark measuring how well AI models and harnesses write fast, accurate, general numerical solvers for large sparse linear systems in C. On the FLASH magnetic-diffusio…
OpenAI launched GPT-6 Sol and GPT-6 Luna, two faster, cheaper models in the GPT-6 family, with API prices cut 50% versus GPT-5.6 promotional pricing — Sol at $2 per million input tokens and $10 output…
OpenAI launched GPT-6 Sol and GPT-6 Luna, two models following GPT-6 Astra, with GPT-6 Sol priced at $2 per million input tokens and $10 per million output tokens — half the previous rate — and GPT-6 …