OpenAI Denies Plans to Release Next-Generation Orion Model This Year
OpenAI officially refutes recent reports claiming its upcoming frontier model, Orion, will launch by December. The company states it has no plans to release an Orion model this year, though it promises other new technology is on the way.
OpenAI officially denies recent reports that it plans to release a next-generation AI model code-named Orion this year. A company spokesperson explicitly tells TechCrunch that there are no plans for an Orion release in 2024, directly contradicting a claim by The Verge that the model would launch by December. However, OpenAI promises that it plans to release a lot of other great technology before the year ends.
The previous reporting suggests that Orion represents a significant step up from OpenAI's current flagship model, GPT-4o, and is partially trained on synthetic data from the company's "reasoning" model, o1. The Verge's original story also claims that close partners, including Microsoft, expect to preview the model as early as November ahead of a broader rollout through ChatGPT. OpenAI previously called that report inaccurate but declines to provide further details at the time.
This carefully worded denial leaves substantial wiggle room regarding OpenAI's actual product roadmap. It remains entirely possible that the company releases a different, less capable model by December, or that the next major model simply does not carry the Orion codename. For now, OpenAI confirms it will continue developing new "GPT" models alongside reasoning models like o1 to address different use cases, leaving the exact timeline for its next frontier model a mystery.