Why OpenAI Indefinitely Delayed Its Open Model

The AI community has been eagerly awaiting OpenAI's next major move, and a powerful, open model was expected to be a highlight of the summer. However, those plans have been put on hold. For the second time, OpenAI is pushing back the release, this time with no new date in sight.

An Indefinite Delay for Safety

OpenAI CEO Sam Altman announced on Friday that the company is delaying the release of its open model indefinitely to conduct more thorough safety evaluations. The model was originally scheduled for release next week after a previous month-long delay.

In a post on X, Altman stated, “We need time to run additional safety tests and review high-risk areas. we are not yet sure how long it will take us.” He emphasized the permanent nature of releasing an open model, adding, “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.”

Aidan Clark, OpenAI’s VP of research leading the project, echoed this sentiment in his own X post, saying, “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.”

The High Stakes of an Open Model

The launch of this open model is a significant event for OpenAI, seen as a crucial step in maintaining its leadership in the AI industry, alongside the anticipated release of GPT-5. Unlike its flagship models, this open version would be freely available for developers to download, modify, and run on their own hardware. This strategy is vital as competitors like xAI, Google DeepMind, and Anthropic continue to invest billions into their own AI development, intensifying the race for dominance.

What We Know About the Awaited Model

Expectations for the model are high. Previous reports indicated that OpenAI aimed for it to be best-in-class among open models, with reasoning capabilities comparable to its proprietary o-series models. The initial delay in June was accompanied by a cryptic comment from Altman about achieving something “unexpected and quite amazing,” which only fueled further speculation. There were also discussions about potentially enabling the model to connect to OpenAI's cloud-based models for handling more complex tasks, though it's unclear if this feature will be included in the final release.

A Competitive Field Heats Up

The delay comes at a time of increasing competition in the open AI model space. Just this week, Chinese AI startup Moonshot AI released Kimi K2, a massive one-trillion-parameter open model that has shown superior performance to OpenAI’s GPT-4.1 on certain benchmarks. This move highlights the rapidly advancing capabilities of competitors and adds pressure on OpenAI to ensure its own open model makes a significant impact upon release. For now, the community must wait as OpenAI takes extra time to address the critical safety and performance standards for its next major offering.