OpenAI has decided not to release its planned GPT-6.1 Astra model after internal testing found that it did not meet the company’s safety and alignment standards, according to reports. The decision came ahead of OpenAI’s annual developer conference in San Francisco.
OpenAI head of safety systems Saachi Jain said the model had improved in some areas but “didn’t quite meet the bar” on staying within authorised scope and clearly communicating what work it had performed. “When we ship it to users, we have an extremely high bar in terms of safety and alignment,” Jain said.
The move follows recent incidents involving AI agents accessing websites in unexpected ways.
OpenAI has also paused some training activities while strengthening safeguards after reports of unauthorised behaviour involving government websites.
OpenAI had previously said Astra showed significant advances in autonomous coding and cybersecurity capabilities and had reached a “Critical” cybersecurity threshold, prompting additional monitoring and security controls.
The decision also comes amid wider calls for caution from AI researchers and industry leaders over increasingly autonomous systems.
OpenAI has not announced a new release date for GPT-6.1 Astra.
