OpenAI, the company behind AI models like ChatGPT, is planning to launch a new AI model called Orion by December. This information comes from a report by The Verge.

Unlike previous models such as GPT-4o, Orion won’t be available on the ChatGPT platform right away. Instead, it will first be offered to partner companies. This allows those companies to create their own products and features using the new model.

Microsoft, which is a major partner for OpenAI, is preparing to host the Orion model on its Azure cloud platform. This hosting could happen as soon as next month.

According to the report, the new Orion model is said to be up to 100 times more powerful than the current GPT-4 model. This makes it a significant upgrade. It is also different from the recently released o1 model, which focuses mainly on reasoning and problem-solving tasks.

OpenAI has been training Orion using synthetic data that comes from the o1 model. This means that the training process involved using artificial data to help improve the model’s performance.

The training for Orion was completed last month, which aligns with a recent post from OpenAI’s CEO, Sam Altman, on social media. In this post, he hinted at the name Orion by mentioning “winter constellations.” Orion is also the name of a well-known star constellation.

Internally, OpenAI believes that Orion is the true successor to the GPT-4 model, although it’s not clear yet if they will officially call it GPT-5.

There is also a possibility that the release date could be pushed to next year, but for now, the company is aiming for a December launch. Overall, this new model is creating a lot of excitement in the tech community.