OpenAI pulls new AI model release following widespread security worries

• OpenAI has been forced to pause its GPT-6.1 rollout amid safety concerns
• Safety head reveals it failed on two separate fronts
• AI developers are calling for more government involvement

OpenAI has reportedly cancelled its planned public release of the upcoming GPT-6.1 Astra model after internal testing revealed potential safety issues.

TheWall Street Journal claims the model was due to be released within Codex and ChatGPT as soon as October 2026, but the company's safety head, Saachi Jain, explained that 6.1 "didn't quite meet the bar."

A modified version of GPT-6.1, or at least a development on GPT-6, will likely still release pretty soon, but the temporary pause highlights ongoing issues that all AI developers seem to be facing with increasingly sophisticated models.

OpenAI halts GPT-6.1 Astra rollout

The WSJ reporting reveals that the model actually failed on two separate instances – firstly, it continued working on a task beyond what a user had initially requested, sometimes taking actions without user permission.

Secondly, 6.1 Astra also failed on the transparency front, apparently showing more deceptive behavior than GPT-6 by obscuring from the user what it had actually done and implying that it had or had not completed certain tasks.

"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," Jain added, citing the company's "extremely high bar in terms of safety and alignment."

The news comes just days after ChatGPT maker OpenAI said in an X post that it would be "conducting a much broader review of actions taken by [its] models during training and evaluation," following a Hugging Face incident.

CEO Sam Altman also asserted that the company is conducting an "extensive and ongoing review."

More broadly, OpenAI and fellow AI developers have also been calling on governments to have more oversight over model releases in order to ensure maximum safety. For now, global ongoings are leading companies to battle it out in a bid to release best-in-class models, but with greater oversight and limitations, they would be freer to slow down and focus more on safety.

TechRadar Pro asked OpenAI to confirm the status of GPT-6.1 Astra and to share further information. Any updates will be provided here.

Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.