Two releases in two days. New: an environment variable that sets which model new sessions start on, an optional spellchecker in the prompt, and a ping when another session goes idle. Plus a sandbox tightening on macOS.
Two weeks ago the message was clear: the intro price ends August 31. Now Anthropic has reversed course - 2 dollars in, 10 dollars out per million tokens stays the standard. Plus the Files API and Admin API go GA.
Anthropic has finished the rollout: Cowork now runs on iPhone, iPad, and the web for every paid plan, not just the priciest tier. The agent that started on the Mac is finally everywhere.
Google is rolling out new learning features across Search and Gemini: interactive visualizations, working 3D simulations, practice quizzes, and a student hub. Gemini Live now keeps researching in the background, too. And students get a free year of AI Pro.
OpenAI is previewing 'Private Safety Processing' - a system that spots misuse patterns across multiple requests without letting staff see the content. The trick: it stays compatible with the Zero Data Retention promise.
On the evening of August 16, a lot of people couldn't get into Claude.ai, Claude Code, and Cowork – the API was hit too. After a good half hour everything was back. Anthropic hasn't named a cause.
The new version resumes on its own after a usage limit, GitLab merge requests get a badge in the footer, and the built-in 'claude-api' skill now costs an eighth of the context it used to. Plus a security follow-up on Windows.
Greg Brockman calls the Hugging Face incident a watershed moment and pushes companies to deploy AI defenders now. New alongside it: Daybreak Blue gives vetted defenders access to GPT-5.6 Sol for security work.