OpenAI is set to unveil a new version of its artificial intelligence model by the end of the year, according to reports. This upcoming iteration, referred to internally as Orion, is anticipated to be far more advanced than the existing GPT-4 model. While a release is targeted for December, it appears that access will initially be limited to select business partners rather than the general public.
OpenAI’s Next Generation AI Model
Technology News has highlighted that OpenAI is aiming for a December 2024 launch for its next large language model. Sources familiar with the development indicated that this new model, previously referred to as Strawberry, is now known as Orion, distinguishing it from the recently developed GPT-4o model.
In the interim, OpenAI has introduced the GPT-4 Turbo and GPT-4o models, but these were not entirely new releases—rather, they were enhancements of the existing GPT-4 framework. The anticipated next model, potentially named GPT-5, promises a complete architectural overhaul and enhanced functionalities.
According to the report, the Orion AI model will not be made publicly available at its launch, marking a change from previous strategies employed by OpenAI. Instead, this model is expected to be provided first to organizations with which OpenAI has established partnerships, allowing these enterprises to develop their own applications before the model is released more broadly.
Microsoft, a significant stakeholder in OpenAI, is likely to be among the initial users of the Orion model. The report also suggests that hosting for the Orion model could begin on Azure servers by November, although the official designation for the new AI model has not yet been confirmed.
Furthermore, the report indicates that Orion could be up to 100 times more powerful than GPT-4 and is expected to offer enhanced AI capabilities. OpenAI’s long-term goal remains the integration of its AI models to achieve artificial general intelligence (AGI).
