OpenAI Dismisses Claim to Release Upgraded AI Model, Orion This Year
Until a day ago, it was assumed that OpenAI intended to launch their upgraded AI model Orion in December.
Now, OpenAI has stated that it has no such intentions for this year.
Altman used X, formerly known as Twitter, to denounce the report as “fake news out of control” within hours of it being published on Thursday night.
However, Altman did not categorically deny any of the numerous aspects in the post. Since Altman didn’t offer a punch list of particular lies, it’s probable that his X post was just a reflexive reaction to the exclusive article.
When the media leaks private information about a significant project, it is not uncommon for business executives to respond in this way.
“We don’t have plans to release a model code-named Orion this year,” a spokesperson told TechCrunch via email. “We do plan to release a lot of other great technology.”
A step ahead of GPT-4o, OpenAI’s Orion frontier AI model was purportedly built in part using synthetic training data from the previously released reasoning model o1.
According to reports, Orion outperforms Project Strawberry, which includes the recently released o1-preview and o1 small.
A number of users have already referred to Orion as possible GPT-5, the next significant advancement in language model evolution that may surpass GPT-4.
Newsng understands that there is a lot of leeway in OpenAI’s statement.
It’s possible that Orion won’t be the company’s next big model.
Or maybe by December, OpenAI will produce a new model, albeit one that isn’t as good as Orion.
We earlier reported that Uber Technologies plans to launch an AI assistant powered by OpenAI technology to aid drivers in shifting to electric vehicles (EVs) as part of its emissions-reduction initiatives.
