OpenAI is preparing to launch its new flagship AI model, Orion, by December 2025. Orion will be the successor to GPT-4, marking OpenAI’s two-year milestone. However, it is unclear whether it will be officially named GPT-5.
Unlike previous models GPT-4o and o1, Orion will not initially be available to the public through ChatGPT. Instead, OpenAI will first offer it to select partner companies, allowing them to integrate the technology before it becomes more widely accessible.
Microsoft to Host Orion on Azure
Sources say that engineers are preparing to host OpenAI’s Orion model on Azure as early as November. OpenAI’s partner Microsoft will lead the deployment, using its Azure cloud infrastructure. Previous reports suggested that Microsoft might acquire OpenAI by 2027 due to OpenAI’s funding challenges.
Read also : OpenAI Secures $6.6B in Funds; Can Decentralized AI Keep Up?
Although Orion is planned for release by December, the timeline could change. OpenAI and Microsoft have not commented on the specifics of Orion’s release.
Orion’s Potential and OpenAI’s Vision
As noted by an OpenAI executive, Orion is expected to have more exponential power than GPT-4, with a potential increase of up to 100 times. This development is reportedly different from the September launch of the o1 reasoning model. The platform’s long-term vision includes integrating its large language models (LLMs), including Orion, to create an ultra-advanced model that approaches Artificial General Intelligence (AGI) capabilities.
In September, OpenAI launched its Strawberry project, called OpenAI o1, with the founding principle of “think before answering.” The company asserted that the AI model employs chain-of-thought reasoning to solve intricate science, technology, engineering, and math (STEM) problems.