OpenAI is reportedly planning to release its next frontier AI model, codenamed Orion inside the company, by December of this year, The Verge reported on Thursday. Unlike previous releases, the company reportedly plans to release the model gradually to trusted partners before a broader rollout through ChatGPT.

An OpenAI spokesperson tells TechCrunch the report is not accurate but would not elaborate further.

The Verge writes that Microsoft engineers expect to receive access to Orion as early as November, although it’s unclear what OpenAI will ultimately call the model. The o1 series of models was codenamed “strawberry” inside OpenAI for months before it was released.

There’s a lot riding on OpenAI’s next frontier model release. The AI startup just raised $6.6 billion in funding at a $157 billion valuation, and investors are expecting OpenAI to continue releasing increasingly capable models to lead the tech world.