ON August 1st, OpenAI released their advanced AI model, Astra, which tackled ten significant mathematical problems across various fields, including high-dimensional geometry and quantum complexity. Unlike previous models, Astra is not publicly accessible for interaction. Instead, OpenAI showcased its capabilities to regulators, emphasizing human-AI collaboration in generating and formalizing mathematical proofs using the Lean 4 theorem prover.
They opened the Lean files to allow researchers to independently verify results. OpenAI's approach aims to outline a new standard for AI in scientific inquiry, highlighting the importance of verification mechanisms as AI's contributions expand beyond simple text and coding.