OpenAI Built GPT-Red to Attack Its Own Models Before Hackers Do
OpenAI built GPT-Red, an internal automated red teamer that attacks AI agents, finds prompt-injection vulnerabilities, and helped harden GPT-5.6 Sol.
Category archive
OpenAI News
Browse the latest posts in this topic and its related archive.
OpenAI Shuts Down ChatGPT Atlas. Browser Ambitions Grow.
OpenAI is shutting down its AI-powered browser Atlas after nine months. But the Chrome extension and cloud desktop browser replacing it represent a bigger…
Fidji Simo Steps Down From OpenAI’s No. 2 Role
OpenAI's No. 2 executive is leaving her full-time role after a prolonged medical leave, leaving Sam Altman without a key lieutenant at a critical…
ChatGPT Work Launches: OpenAI’s Multi-Hour Agent With Codex and GPT-5.6
OpenAI launched ChatGPT Work, a multi-hour agent that runs across apps and files. It's powered by Codex, GPT-5.6, and lands two days after Claude…
GPT-5.6 Goes GA: Sam Altman Calls It OpenAI’s Best Model Ever as Sol, Terra, and Luna Launch Worldwide
OpenAI launched the GPT-5.6 family to general availability on July 9 with Sam Altman calling it their best model ever. Sol, Terra, Luna, and…
OpenAI Just Found 30% of SWE-Bench Pro Tasks Are Broken, Retracting Its Recommendation
OpenAI audited SWE-Bench Pro and estimates about 30% of its tasks are broken. The company is retracting its recommendation that the research community use…
OpenAI Fixed an 18-Year-Old Bug in a Library Everyone Uses. Here’s How They Found It.
OpenAI engineers debugged a years worth of data infrastructure crashes and found two bugs - one in bad hardware and another unnoticed in open…
Codex Gets a UX Update: Smoother Long Threads, a Hoverable Nav Rail, and Better Settings Search
Codex UX improvements landed: smoother long thread handling, a hoverable nav rail, and expanded settings search with host-filtering controls.
OpenAI Just Made Its Own AI Chip. Here’s What Jalapeño Actually Means.
OpenAI and Broadcom unveiled Jalapeño, a custom LLM inference chip built from the ground up. Early testing shows performance per watt substantially better than…
A Mother Is Suing OpenAI After Her Daughter’s Suicide. ChatGPT Kept Talking.
A Canadian mother is suing OpenAI after her daughter's suicide. The lawsuit claims ChatGPT evolved from redirecting to crisis hotlines into a fake confidant…
Codex Finally Lets You Bank Your Rate Limit Resets Instead of Losing Them
OpenAI just announced that Codex users can now save rate limit resets for later instead of losing them on OpenAI's schedule. If you've ever…
OpenAI Just Bought the Missing Piece for AI Agents That Work While You Sleep
OpenAI just announced it's acquiring Ona, and if you only read the headline, you might think this is just another acqui-hire. It's not. When…