- OpenAI has confirmed it will not release an AI model codenamed “Orion” this year.
- The Verge reported that Orion, OpenAI’s anticipated next model, could launch by December, but OpenAI disputed this claim.
OpenAI has clarified that it has no intention of releasing an AI model codenamed “Orion” this year, contradicting recent speculations about its product roadmap.
“We don’t have plans to release a model code-named Orion this year,” an OpenAI spokesperson confirmed to TechCrunch via email. “We do plan to release a lot of other great technology.”
This statement directly addresses a report from The Verge, which recently claimed that OpenAI would launch Orion by December, with exclusive previews for trusted partners like Microsoft, who may access it as early as November. Microsoft, an OpenAI partner and investor, is reportedly anticipating access to this advanced model soon.
Orion, rumored to be a leap forward from OpenAI’s current GPT-4 model, is said to be partially trained on synthetic data generated by OpenAI’s o1 model, which specializes in “reasoning” tasks. Unlike GPT-4, Orion is expected to focus on different applications, potentially establishing a separate line of AI advancements alongside the ongoing GPT developments. This roadmap signals OpenAI’s commitment to developing both advanced GPT models and reasoning models for varied use cases.
OpenAI’s careful phrasing leaves open the possibility that another model may still debut this year, although perhaps not as sophisticated as Orion. Whether Orion or an entirely different model emerges remains to be seen, adding intrigue to OpenAI’s development trajectory as the year progresses.
Edited by Harshajit Sarmah
