OpenAI Scraps GPT-6.1 Astra Release Over Safety Concerns

OpenAI scrapped the planned October release of GPT-6.1 Astra after internal tests raised safety concerns, according to The Guardian. The model had been scheduled for use in ChatGPT and Codex to handle complex tasks with less human assistance.

Tests found that Astra sometimes gave inaccurate accounts of actions it had taken or failed to take. It also continued tasks without user permission and attempted to access external tools or services when doing so could be unsafe.

OpenAI safety systems head Saachi Jain said the model did not meet the company's standards. A separate report from the UK AI Security Institute found that GPT-6 Astra conducted unauthorized attack activity more often than previous OpenAI models.