Well, OpenAI is preparing to launch its next frontier AI model, codenamed Orion, by December. OpenAI released ChatGPT as a research preview on November 30, 2022.
Sources familiar with the plan told The Verge the company is taking a different approach to this release, initially restricting access to select corporate partners rather than making it widely available through ChatGPT.
Microsoft, OpenAI’s primary partner, is reportedly preparing to host Orion on its Azure platform as early as November. While the model is considered the successor to GPT-4 internally, OpenAI has not confirmed whether it will be marketed as GPT-5.
The new model has generated significant hype, with one OpenAI executive suggesting it could be up to 100 times more powerful than GPT-4. Unlike the recently released o1 reasoning model, Orion will likely represent a major advancement in the company’s core technology. Sources indicate that OpenAI has been using o1 (codenamed Strawberry) to generate synthetic training data for Orion.
The timing of Orion’s possible release aligns with major developments at OpenAI, including a recent US$6.6 billion funding round requiring the company’s restructuring as a for-profit entity.
The company also went through a recent exodus of executives with the departures of CTO Mira Murati, Chief Research Officer Bob McGrew, and VP of Post Training Barret Zoph, to name a few.
Neither OpenAI nor Microsoft have released any statements on Orion’s planned release.
