OpenAI is preparing to launch its newest artificial intelligence model, known internally as Orion, by December. The company is taking a different approach with this release compared to its previous models, opting for a controlled rollout that begins with select business partners.
According to information obtained by The Verge, the initial release will not be available through ChatGPT, the company’s popular AI chatbot platform. Instead, OpenAI plans to provide access first to specific companies it collaborates with, allowing them to develop their own products and features using the new model.
Microsoft Integration
Microsoft, OpenAI’s primary partner for AI model deployment, is already making preparations for Orion. Engineers at Microsoft are working to host the new model on their Azure cloud platform, with preparations potentially beginning as early as November. While Orion is considered the successor to GPT-4 within OpenAI, the company has not yet confirmed whether it will be marketed as GPT-5.
Development and Company Changes
According to internal sources the development of Orion reached a significant milestone in September when OpenAI researchers celebrated the completion of the model’s training phase. This timing coincided with a message posted on X by OpenAI CEO Sam Altman, who made a reference to winter constellations, apparently alluding to the project’s codename.
The launch of Orion comes during a period of significant changes at OpenAI. The company recently secured a $6.6 billion funding round and is undergoing restructuring into a for-profit entity. Additionally, the organization is experiencing leadership changes, with the departures of Chief Technology Officer Mira Murati, Chief Research Officer Bob McGrew, and Vice President of Post Training Barret Zoph.
When approached for comment about the reported December release timeline, OpenAI representatives declined to provide additional information. A spokesperson later told TechCrunch that the reported information was not accurate, though no further details were provided.
The launch of a new frontier model will be a crucial development for OpenAI, particularly following its valuation at $157 billion. Investors are looking to OpenAI to maintain its position at the forefront of artificial intelligence development through the release of increasingly advanced models in the face of fierce competition.
Rival Anthropic recently introduced the latest advances in its Claude AI. The Claude 3.5 lineup showcased two updated models that target different performance, speed, and cost brackets.
The flagman Claude 3.5 Sonnet features an advanced “Computer Use” mode which allows the model to execute commands on desktop applications, from web browsing to complex app control. It enhances coding capabilities, outperforming even leading models like GPT-4 in specific benchmarks. The Claude 3.5 Haiku is all about efficiency and cost-effectiveness. It caters to simpler, high-speed interactions, providing responsive AI support suitable for customer service and real-time applications.
Apart from these two major players there are also the open-source LLMs by Meta, called LLAMA and even Nvidia has entered the race by dropping a fine-tuned LLAMA model that beats ChatGPT and other models on certain benchmarks.
