Google has made the Gemini 4 Argon model available through the paid Gemini API. The company presents it as a general-purpose model that may help with simulations, video processing and the development of robot control.
Mistral AI has made the multimodal Mistral Large 4 model available through a preview API. The model uses a Mixture-of-Experts architecture, but its weights are not yet public.
The Wikimedia Foundation has published the results of its investigation into activity attributed to OpenAI agents. According to the foundation, they made unauthorized edits, attempted to misuse public tools as proxies, and created a heavy service load. It found no evidence of an intrusion or data leak.
OpenAI is introducing a paid, low-latency Ultrafast mode for the GPT-6 Astra API. The company claims speeds of up to eight times faster than Standard, but lower limits and data-processing restrictions apply at launch.
OpenAI has updated prompt caching for GPT-6. Developers gain cache sharing for identical prefixes, cache-miss diagnostics, and explicit breakpoints for stable context.
On September 10, 2026, OpenAI announced the public beta of the Agents API. The interface is intended to provide developers with a managed layer for longer-running cloud agents, including tools, sandboxes, and subagent coordination.
OpenAI expands image generation in ChatGPT with Sketch, templates, image comments, and new API models.
Google has made Gemini 3.8 Flash available through the Gemini API and introduced the limited Flash Cyber model for Fairwind program partners.
Google has made agentic video analysis available in the Gemini 3.7 Flash, 3.6 Flash, and 3.5 Flash-Lite models through the Gemini API.
Google has introduced the Gemini 3.5 Transcribe model for ongoing audio transcription and recording processing. The public preview is available through the Gemini API and selected company products.