OpenAI News.
4 stories in the last 7 days
Everything OpenAI across dozens of tech sources, distilled by AI into sharp ~100-word summaries and updated hourly. Follow OpenAI to pin its stories to your For You feed. Tap any story for the full brief, or open the original source.
Anthropic announced support for OpenAI's AGENTS.md specification in Claude Code, allowing the tool to use the rival's markdown instructions file. This move eases compatibility for developers using both platforms.
A three-person security team used Anthropic's Claude Opus 5 to breach OpenAI's internal systems, including employee accounts and its GitHub repository, in under 72 hours. The attack cost less than $3,000 in AI tokens.
OpenAI released GPT-6 Astra, an AI model that controls computers and apps. The launch video showed it using Blender, prompting immediate backlash from the Blender and game-developer communities over concerns about art piracy and human creativity.
A three-person team from Hackron AI used Anthropic's Claude Opus 5 to breach OpenAI's internal systems on July 25, exploiting a Discourse forum vulnerability to access employee accounts and a private GitHub monorepo.
Summaries by ByteBrief