OpenAI apparently plans to release Orion, its next major AI model, by the last month of 2024, December.
According to The Verge, unlike the last two OpenAI models, GPT-4o and o1, Orion’s AI will not initially be released through ChatGPT. Instead, OpenAI plans to make its product available first to companies it works closely with so they can build their own products and features.
Microsoft’s readiness to host Orion artificial intelligence
Another source told The Verge that Microsoft engineers (OpenAI’s main partner for deploying artificial intelligence models) are preparing to host Orion on Azure in early November. While Orion appears to be touted as a successor to GPT-4, it’s unclear whether OpenAI will call it GPT-5.
Earlier, an OpenAI executive at X claimed that Orion was potentially 100 times more powerful than GPT-4. Another report also claimed that OpenAI was using o1 codenamed “Strawberry” to train Orion.
However, an OpenAI spokesperson told TechCrunch that the Verge report is not accurate, so the Orion AI model may be announced at a later date.
Reports of the release of OpenAI’s next big model come at a critical time for the startup, which just raised $6.6 billion in funding. The startup is also changing its internal team, with OpenAI’s CTO Mira Moratti recently announcing her resignation, along with other members including Chief Research Officer Bob McGraw and VP of Education Bart Zoff. .
RCO NEWS