OpenAI to launch next AI model ‘Orion’ by December: What to expect

OpenAI is gearing up for the release of its next-generation AI model, codenamed Orion, with a potential launch as early as December 2024. Unlike previous model rollouts, which were immediately available on ChatGPT, OpenAI is taking a more cautious approach this time by initially offering access to trusted partners. These companies will be able to build custom products and features using the cutting-edge technology, marking a strategic shift in how the AI giant rolls out its models.

Sources suggest that Orion, said to be up to 100 times more powerful than GPT-4, will be hosted on Microsoft’s Azure platform as soon as November. Microsoft, a key OpenAI partner, is already preparing for its integration, showcasing the ongoing collaboration between the two tech powerhouses. However, despite the heightened anticipation, neither OpenAI nor Microsoft has officially confirmed the launch timeline.

A new approach to AI deployment

Unlike the more public rollouts of previous models like GPT-4, OpenAI’s strategy with Orion revolves around a targeted release to a select group of partners. This will allow companies with close ties to OpenAI to test the model’s capabilities and build bespoke products tailored to their specific needs. This calculated approach aims to ensure that Orion delivers meaningful value to key industries before a wider public release.

This shift also aligns with OpenAI’s long-term vision of building toward artificial general intelligence (AGI). The company’s leadership has long expressed ambitions to combine large language models (LLMs) to achieve a broader, more capable AI system. Orion represents a significant step toward this goal, laying the groundwork for more advanced applications of AI in various sectors.

Anticipation builds amidst a crucial financial milestone

The timing of Orion’s release is critical. OpenAI recently raised $6.6 billion in funding, which has driven the company to restructure as a for-profit entity. Investors have high expectations for the tech leader to continue releasing increasingly sophisticated models. With its enhanced capabilities, Orion is expected to maintain OpenAI’s position at the forefront of AI innovation.

Orion’s development process has been equally ambitious. Reports indicate that the model was trained using synthetic data generated by the o1 reasoning model, known internally as “Strawberry.” OpenAI engineers completed training in September, and CEO Sam Altman has since teased the model’s arrival with cryptic social media posts referencing the Orion constellation, sparking speculation across the tech community.

As Orion approaches its launch, excitement is building within both the AI and business sectors. If the model lives up to its promise, it could set a new benchmark for artificial intelligence, driving innovation across industries and inching OpenAI closer to its AGI vision.

• Read more under:
• Announcements
• Artificial Intelligence