OpenAI is preparing to release a new artificial intelligence model called Orion by the end of 2024. The president of the company's Japanese division said that the new model could be 100 times more powerful than GPT-4, The Verge reports.
'GPT Next’ to Achieve 3 OOMs Boost. Great insights from the #KDDISummit. Tadao Nagasaki of @OpenAI Japan unveiled plans for ‘GPT Next,’ promising an Orders of Magnitude (OOMs) leap. ⚡️ This AI model aims for 100x more computational volume than GPT-4, using similar resources but… pic.twitter.com/fMopHeW5ww
— Shaun Ralston (@shaunralston) September 3, 2024
Orion will not be immediately available to ChatGPT users. Instead, OpenAI will give access to the new model to its closest partners so that they can develop new features and products based on it.
According to preliminary reports, the new model is scheduled to be released in December 2024, and Microsot is preparing to adopt Orion on Azure servers in November. However, neither company has provided official comments on this topic.
OpenAI used the recently introduced o1 model, which can "think", to provide Orion with synthetic data for training. In the future, the company aims to combine large language models to eventually create a general artificial intelligence. Among other things, it could help OpenAI exit its partnership with Microsoft.