OpenAI is set to release its next-generation artificial intelligence model, codenamed Orion, by December 2023. This launch marks nearly two years since the debut of ChatGPT, OpenAI's popular AI chatbot.
Unlike previous models, OpenAI plans a phased release for Orion, initially limiting access to its close corporate partners. These firms will be able to build and test customized applications using the model. Microsoft, OpenAI's primary deployment partner, is preparing to host Orion on its Azure cloud platform as early as November, although specific details about the release timeline remain unconfirmed.
Orion is expected to surpass the capabilities of its predecessor, GPT-4, with an OpenAI executive suggesting it could be up to 100 times more powerful. The model's development has utilized synthetic data generated through the recently launched o1 model, enhancing Orion's training process.
OpenAI's CEO, Sam Altman, hinted at the release in a social media post, referencing the visibility of the Orion constellation in the northern hemisphere starting in November.
The timing of Orion's release is notable, following a $6.6 billion funding round that necessitated OpenAI's transition to a for-profit structure. This shift has resulted in significant internal changes, including the departures of key executives.