FrontierAi2 open-sources AstaBrief 8B, the model behind Asta's Fast mode: a cited report about 3.5 times faster than the Claude mode, per Ai2Ai2 has published the model and its training data. From a research question and excerpts from the literature, AstaBrief writes a cited report, and in Asta it runs as Fast mode next to the Claude-powered Thinking mode. Ai2 itself says the training and evaluation were done mostly in 2025 and that the full evaluation has not been rerun against today's leading models.FrontierAi2 releases Olmo-core 3: open code for training MoE models up to a trillion parametersThe Allen Institute has shown a rebuilt training system for mixture-of-experts models. Per Ai2, on 8 NVIDIA B300 chips a 47 billion parameter model trains 2.7 times faster than with their old code, and the largest test run passes 1.2 trillion parameters on 512 GPUs. They ship a report, code and a demo. No models.FrontierGoogle launches Guided Vision in Gemini Live: the camera says out loud what is in front of a blind userYou point the phone, Gemini tells you what it sees and how to move the frame. The feature is for blind and low-vision people and works on Android 9 and up wherever Gemini Live is available. Google states plainly that it is not a medical device and does not replace the white cane.FrontierGPT-6 Astra now has an Ultrafast mode: up to 8x faster generation, per NVIDIAThe mode runs on NVIDIA Blackwell and is available in the OpenAI API and to eligible ChatGPT Work and Codex users. OpenAI has no announcement of its own in its newsroom; the number is in NVIDIA's blog, with quotes from two OpenAI people, and in OpenAI's API documentation. OpenAI's documentation adds the condition that, for us, comes first: US or global processing only, no EU endpoints.FrontierGoogle releases Gemini 4 Argon to cyber defenders first, and without cyber brakesOn 30 September Google announced Gemini 4 Argon, its new top model. It is going to trusted defenders through the Fairwind programme, and for them and Google's internal teams it will come without cyber safeguards. The broad release is a plan: paid API customers and Google AI Ultra subscribers first.FrontierOpenAI releases GPT-6.1 Sol: nearly Astra-level on agentic work, at a fifth of the priceOn 29 September OpenAI released GPT-6.1 Sol, an upgrade to GPT-6 Sol. Per the company it nearly matches GPT-6 Astra on agentic coding, computer use and professional tasks at a fifth of Astra's token prices. It is available in ChatGPT Work, Codex and the API, but not yet in regular chat.FrontierClaude Sonnet 5.5 ships at the old price but typically gets the job done with fewer tokensOn 28 September Anthropic released Sonnet 5.5, the second model in the 5.5 family. Per the company it is over 30 percent faster than Sonnet 5 and up to 30 percent cheaper per task, at the same 2 and 10 dollars per million tokens. It is also the first Sonnet to launch with cyber safeguards.FrontierAlibaba is training Qwen 4 and plans its first cloud regions in Finland and the NetherlandsAt the Apsara 2026 conference Alibaba said the next Qwen generation is in training, and the Qwen 4.5 and Qwen 5 series are planned to reach 5 to 10 trillion parameters. Its cloud plans new regions in Türkiye, Finland and the Netherlands in the next 12 months. Everything with a future date is a plan.
80 stories