GPT-4's successor is scheduled to appear in December – which contradicts the head of OpenAI

With the release of the next major language model from OpenAI, things may be different than usual. As US tech magazine The Verge reported, citing anonymous sources at the company, a successor to GPT-4 could be launched in December. OpenAI President Sam Altman Report called However, in the social network

The new Large Language Model (LLM), developed under the code name Orion Edge According to available information, it will not be initially usable via ChatGPT. Instead, OpenAI plans to initially grant access to close partners. They should be able to use the advantage of time to adapt or expand their products based on the new model. According to internal schedules, Microsoft expects to host Orion on the Azure cloud platform starting in November.

100 times stronger than GPT-4?

Orion is said to be 100 times more powerful than GPT-4, although it may not be called GPT-5. The training was completed in September. This year, OpenAI released an improved version of GPT-4 called GPT-4o in May and OpenAI o1 as a specific extension in September. With the December release, OpenAI supposedly wanted to celebrate the second anniversary of the ChatGPT release.

The report criticized by Altman goes on to say that OpenAI used its AI model o1 to provide synthetic data to train Orion. Altman also made a reference to Orion in a mysterious X-post post in which he explained that he's excited about the winter constellations that will soon rise — after all, the namesake constellation is best visible from November to February.

(Makki)