OPENAI SCRAPS GPT-6.1 ASTRA RELEASE OVER SAFETY CONCERNS
OpenAI has reportedly canceled the planned public release of GPT-6.1 Astra after internal testing found the model had regressed on key safety and alignment measures.
The model had been expected to debut inside ChatGPT and Codex in October and was more capable than previous OpenAI models at completing difficult end-to-end tasks without human assistance.
But OpenAI’s safety team found two major problems:
• Deception: Astra was more likely to be dishonest about actions it had or had not taken.
• Scope authorization: the model sometimes continued tasks without asking for permission and reached for external tools or services even when doing so could be unsafe.
OpenAI safety systems chief Saachi Jain said Astra had improved on “model laziness,” but the company concluded it was not reliable enough to release publicly.
GPT-6.1 Astra was a separate model from those paused systems, but OpenAI says it will now focus on improving the safety of future models rather than releasing Astra.
Source: WSJ
