OpenAI is preparing a notable shift in how it releases AI models. The company says it intends to publish its first open language model since GPT‑2 in the coming months, and it is asking the wider AI community what that model should offer.
The plan was disclosed through a feedback form OpenAI published on its website Monday. The form is aimed at developers, researchers, and members of the broader community, signaling that OpenAI wants outside input before the model arrives.
OpenAI is asking what an open model should include
The feedback form centers on practical questions. OpenAI asks what people would like to see in an open-weight model from the company and what open models they have used before.
That framing matters because it places the release in a different category from OpenAI's familiar hosted products. An open-weight model is expected to be something the community can examine and use more directly than a model available only through a controlled service.
OpenAI described the process as a collaboration with developers, researchers, and the broader community. The company wrote that it wants to gather input and make the model as useful as possible, and it invited people who are interested to join feedback sessions with the OpenAI team.
Steven Heidel also posted on X on March 31, 2025: "we’re releasing a model this year that you can run on your own hardware". That statement points to one of the central questions around the release: who will want to run the model themselves, and what will they build with it?
Developer events will shape the rollout
OpenAI plans to host developer events as part of the feedback process. Those events are meant to gather input and, in the future, show prototypes of the model.
The first developer event will take place in San Francisco within a few weeks. After that, OpenAI plans sessions in Europe and Asia-Pacific regions.
This sequence suggests a staged approach. OpenAI is not simply announcing a finished model and moving on; it is creating a channel for developers and researchers to comment before the final release. The company is also setting expectations that prototypes may be shown later, which gives the community a clearer path from feedback to demonstration.
For users of open models, the details will matter. The source does not specify the model's size, license terms, training data, or release date beyond the coming months. But OpenAI's questions make clear that past experience with open models is part of what it wants to understand before launch.
Rival pressure is part of the backdrop
The move comes as OpenAI faces increasing pressure from rivals that have taken a more open approach to AI model launches. The source identifies Chinese AI lab DeepSeek as one of those competitors.
Unlike OpenAI's strategy, these open competitors make models available to the AI community for experimentation and, in some cases, commercialization. That approach has become a meaningful competitive force.
Meta is another example named in the source. The company has invested heavily in its Llama family of open AI models, and earlier in March it said Llama had racked up over 1 billion downloads.
DeepSeek has also built momentum. The source says it has quickly amassed a large worldwide user base and attracted the attention of domestic investors.
Taken together, those examples show why OpenAI's open language model plan is more than a product update. It is also a response to a market where open AI models have become a serious route to developer adoption, experimentation, and broader distribution.
Sam Altman has signaled a strategy change
OpenAI CEO Sam Altman has also publicly acknowledged the debate over openness. In a recent Reddit Q&A, he said he thinks OpenAI has been on the wrong side of history when it comes to open sourcing its technologies.
"[I personally think we need to] figure out a different open source strategy," Altman said. "Not everyone at OpenAI shares this view, and it’s also not our current highest priority […] We will produce better models [going forward], but we will maintain less of a lead than we did in previous years."
That comment is important because it separates two ideas. OpenAI still expects to produce better models going forward, according to Altman, but he also said the company will maintain less of a lead than it did in previous years.
In that context, an open model can serve several purposes. It can give developers something they can run and modify. It can give researchers a more direct way to experiment. It can also give large companies and governments an option where they prefer to operate a model themselves.
The model is expected to include reasoning capabilities
Altman expanded on the plan in a post on X on Monday afternoon. He said the upcoming open model will have reasoning capabilities along the lines of OpenAI's o3-mini.
He also said OpenAI will evaluate the model according to its preparedness framework before release, as it would for any other model. Because the model can be modified after release, Altman said OpenAI will do extra work.
The safety and review context is especially visible because excerpts of a forthcoming book by Wall Street Journal reporter Keach Hagey, published over the weekend, allege that Altman misled OpenAI executives about model safety reviews prior to his brief ouster in November 2023.
For now, the central facts are clear but limited. OpenAI says an open language model is coming in the coming months, it is asking the community for input, and it plans developer events before broader release. The unresolved details are the ones developers will be watching most closely: what the model can do, how it can be used, and how OpenAI balances openness with its stated review process.