OpenAI drops another batch of mathematical breakthroughs

OpenAI has released a new batch of mathematical breakthroughs, including solutions to hundreds of open questions. The company had previously announced in September that its model had "resolved more than 100 long-standing open problems across most areas of mathematics." The newly released manuscripts provide details on the problems solved, summaries of the model's reasoning, compute usage estimates, and statistics on attempted problems. OpenAI claims the "average result" utilized the equivalent of three hours of ChatGPT Pro thinking.

This release provides the first details on which problems OpenAI's model solved, unlike earlier announcements that only disclosed the number of problems.