OpenAI has scrapped the release of GPT-6.1 Astra after internal testing found the system did not meet the company’s safety and alignment standards, the ChatGPT maker has confirmed.
The next-generation AI model had been planned for an October debut.
OpenAI chief executive Sam Altman and Dario Amodei, head of Anthropic, earlier this month joined industry leaders in calling for a slower pace of AI development and stronger safety measures.
OpenAI has warned that Astra, its flagship GPT-6 model, can at times evade human oversight, while the company and rivals such as Anthropic have faced scrutiny over experimental AI systems that breached safeguards, including an OpenAI model that accessed Australia’s health system database.
READ MORE
The Wall Street Journal reported earlier in the day that OpenAI had abandoned plans to launch the model, which was expected to be integrated into ChatGPT and Codex and was designed to handle more complex tasks without human assistance.
The WSJ reported that GPT-6.1 Astra also showed higher levels of deception than its predecessor in internal testing, including instances in which it did not always accurately disclose what actions it had taken.
[ OpenAI systems go rogue and meddle with US state sitesOpens in new window ]
“While [GPT-6.1 Astra] improved on axes such as model laziness, it didn’t quite meet the bar in terms of staying within scope and authorisation, and how it communicates back to the user about the type of work it’s done,” said Saachi Jain, head of safety systems at OpenAI.
“Of course we want to make sure our model development is safe no matter whether that’s in the company or when we ship it to users. But when we ship it to users, we have an extremely high bar in terms of safety and alignment,” Jain said.
The decision comes in advance of OpenAI’s developer conference in San Francisco, where the company has previously unveiled products aimed at software developers.
[ OpenAI boss ‘confident’ industry can police AI safetyOpens in new window ]
Meanwhile, OpenAI on Tuesday apologised for its AI models breaching Australian government websites and promised to set up a new task taskforce as part of reforms after the incident.
[ AI agents are hacking government agencies. It’s ‘really scary’ for cybersecurityOpens in new window ]
The San Francisco-based company said it would fund cyber defences in Australia and help manage the risk of increasingly capable artificial intelligence, as it navigates the fallout after its models accessed Australian websites without authorisation. It gave some of the details of the June incident and its intended steps to avoid a repetition and restore trust in a blog post.
“We are sorry and working to do better in the future,” OpenAI wrote. “People want to know AI is being developed safely, and that starts with what companies like ours do ourselves.” – Reuters/Bloomberg
(c) Copyright Thomson Reuters 2026
