![]()
OpenAI plans to release an “open” version of its language model later this year, enabling developers to run the model on their own hardware.
On March 31, CEO Sam Altman posted an update on X explaining that the company will soon introduce a powerful “new open-weight language model with reasoning.” However, before the launch, they aim to collect feedback on how to optimize its utility.
![]()
Altman mentioned, “We’ve been considering this for a long time, but other priorities came first. Now, it seems like the right time to proceed,” noting that this marks the first “open-weight” release since GPT-2 in 2019.
He added, “We still have some decisions to make, so we are organizing developer events to gather input and eventually experiment with early prototypes.”
Unlike fully closed models like GPT-3 and GPT-4, an open-weight model is accessible for anyone to use, download, modify, or deploy for their own needs, although it is not as unrestricted as an open-source model.
Developer events for this model will commence in San Francisco—where OpenAI is based—in the coming weeks, followed by sessions in Europe and the Asia–Pacific region.
![]()
Altman expressed enthusiasm about the potential, saying, “We’re excited to see what developers build and how large companies and governments utilize it, especially if they prefer to run a model in-house.”
He also noted, “We will put in extra work knowing that this model might be modified after its release.”