Rumor has it that OpenAI is planning to release a cutting-edge model codenamed Orion before December, but OpenAI reportedly has no intention of making Orion widely available to users. Instead, it will first license the model to companies closely partnered with OpenAI to develop products and new features, and only later consider opening it up to the general public.
OpenAI’s next-generation model Orion expected to debut in December: performance could be 100 times that of GPT-4
According toforeign mediaSources say OpenAI plans to launch its next-generation frontier model Orion by December this year. However, unlike GPT-4o and the o1 model, Orion will be licensed to companies that work closely with OpenAI, allowing them to develop their own products and features. One person familiar with the matter revealed that engineers at Microsoft, OpenAI’s backer, are planning to deploy Orion on Azure as early as November.
Regarding Orion’s performance, OpenAI executive @shaunralston revealed on X that Orion could be 100 times more powerful than GPT-4. Additionally, unlike the o1 reasoning model OpenAI released in September, OpenAI’s goal for Orion is to integrate its LLM models over time to create a more powerful model, which may eventually be called Artificial General Intelligence (AGI).
‘GPT Next’ to Achieve 3 OOMs Boost. Great insights from the #KDDISummit. Tadao Nagasaki of @OpenAI Japan unveiled plans for ‘GPT Next,’ promising an Orders of Magnitude (OOMs) leap. ⚡️ This AI model aims for 100x more computational volume than GPT-4, using similar resources but… pic.twitter.com/fMopHeW5ww
— Shaun Ralston (@shaunralston) September 3, 2024
Already hadreportindicating that OpenAI is using the codename Strawberry The o1 model provided synthetic data for Orion’s training, and sources familiar with the matter also revealed that OpenAI researchers had already held a celebration party in September to celebrate the successful training of the new model.
Although OpenAI has recently experienced turmoil such as executive departures and the exclusive funding agreement, OpenAI has consistently been able to prove that it is doing well. Although the text-to-video generation model Sora has been talked about for a long time and still hasn’t been released, OpenAI continues to update and release models with different use cases. If you’re interested in OpenAI’s recent turmoil and the models it has released, you can visit the link below to read related coverage:
Source: KOCPC Chinese