OpenAI shelves GPT-6.1 Astra over safety tests

OpenAI will not ship GPT-6.1 Astra, a next-generation agent model planned for an October debut in ChatGPT and Codex, after internal tests found it fell short of the company's safety and alignment bar, the company confirmed on September 28. Head of safety systems Saachi Jain said the model improved on laziness but did not meet the bar on staying within scope and authorization or on how it reports the work it has done; the Wall Street Journal first reported the decision, and Reuters, CNBC and the BBC confirmed it with the company. This is a product-release hold, distinct from the September 20 DNS sandbox training pause already on the OpenAI file; GPT-6 Astra, Sol and Luna remain available, and DevDay is still scheduled for September 29.