- OpenAI is planning to release its next big model, GPT-5 aka Orion by December this year.
- Microsoft is preparing to host the Orion model on its Azure infrastructure by November.
- The frontier model will be first available to close partners of OpenAI to develop new features and experiences.
It has been over a year and a half since OpenAI released GPT-4, yet the company has not released a big frontier model. Sure, GPT-4o and OpenAI o1 models were launched with new upgrades, but users have been asking for GPT-5 with advanced capabilities. Now, a report by The Verge reveals that OpenAI may release GPT-5, codenamed ‘Orion’ by December this year.
The report further says that Microsoft is preparing to host the Orion model on Azure by November. OpenAI will be releasing the model to close partners first to deliver new features and products, likely via the API. That said, reacting to the report, OpenAI CEO Sam Altman responded on X, “fake news out of control.“
In May 2024, Microsoft CTO Kevin Scott presented a graph showing upcoming OpenAI GPT models will scale tremendously and require massive computing resources.
In September, Sam Altman shared a cryptic post on X, “excited for the winter constellations to rise soon; they are so great.” If you ask ChatGPT what this tweet alludes to, it reveals ‘Orion’ which appears during the winter months of December and January. Altman is likely referring to the upcoming Orion aka GPT-5 model.
Whatever the case, it seems OpenAI is gearing up to launch its next big model soon. Anthropic recently upgraded the Claude 3.5 Sonnet model which gets even better at coding and other tasks. Amid many researchers and executives leaving OpenAI, the company is in a tight spot to keep up with the momentum.
