OpenAI has once again delayed the release of its highly anticipated open AI model, citing the need for further safety testing.

The move comes just weeks after the company had already pushed back the release schedule, originally set for earlier this summer.

The delay was confirmed by OpenAI CEO Sam Altman in a post on X (formerly Twitter) on Friday, where he said the model would not be released next week as previously expected. “We need time to run additional safety tests and review high-risk areas. We are not yet sure how long it will take us,” Altman stated. “While we trust the community will build great things with this model, once weights are out, they can’t be pulled back. This is new for us and we want to get it right.”

The upcoming release was to mark OpenAI’s first openly accessible AI model in years and was expected to be available for developers to download and run locally — a significant departure from the company’s usual proprietary approach.

The delay comes at a time when competition in the open-source AI landscape is heating up. On the same day, Chinese AI startup Moonshot AI announced the launch of its Kimi K2 model — a one-trillion-parameter system that reportedly outperforms OpenAI’s GPT-4.1 on several advanced coding benchmarks.

OpenAI’s Vice President of Research, Aidan Clark, who is overseeing the open model team, also commented on the delay. “Capability wise, we think the model is phenomenal — but our bar for an open source model is high and we think we need some more time to make sure we’re releasing a model we’re proud of along every axis,” Clark posted on X.

Industry watchers had expected OpenAI’s open model to be on par with its o-series models in terms of reasoning abilities. TechCrunch had earlier reported that the company was aiming to deliver a best-in-class open model, further raising expectations among developers and researchers.

While the delay is unlikely to diminish OpenAI’s stature in the field of artificial intelligence, it does underscore the challenges even leading labs face in balancing innovation with ethical and safety concerns. The AI ecosystem has become increasingly crowded in recent months, with firms like xAI, Google DeepMind, and Anthropic all investing heavily in next-generation models.

In June, Altman hinted at the significance of the new model during the announcement of its initial delay, stating the team had achieved something “unexpected and quite amazing,” though he stopped short of offering specifics.