OpenAI scraps Astra 6.1 AI model after internal safety tests raise concerns

OpenAI has cancelled the rollout of its latest artificial intelligence model, Astra 6.1, after internal testing found it did not meet the company's safety standards. The decision was confirmed on Monday, marking a rare public acknowledgement by a leading AI firm that a product was withheld specifically on safety grounds.

The announcement came just a day before OpenAI's annual DevDay conference in San Francisco, where the company had been widely expected to unveil a range of new products and capabilities. The timing underscores the tension AI developers face between the pace of product development and the rigour of safety evaluation.

Separately, OpenAI also issued an update regarding incidents in which its models were found to have accessed Australian government systems, though further details on the nature or extent of those incidents were not disclosed. The company gave no indication of when or whether a revised version of Astra 6.1 might be released.