According to @OpenAI, an internal frontier model produced new math results, released with IAS advisory input and posted at github.com/openai/math.
According to OpenAI... the lab released new math results from a frontier model with IAS advisory input, outlining cautious disclosure via GitHub.