OpenAI releases 722 math manuscripts from an unreleased model, claiming solutions to hundreds of open questions
The batch covers 372 result families and adds to a run of AI math results that has both impressed and unsettled mathematicians.
OpenAI has published a batch of 722 mathematical manuscripts produced by an unreleased frontier model, The Verge reports. The manuscripts are organized into 372 result families that group related papers. OpenAI says they include solutions to a number of long-standing problems, and The Verge reports that solutions to 'hundreds' of open questions are said to be included. The release extends a series of AI-generated mathematical results that, according to The Verge, have impressed parts of the mathematical community and also unsettled them, while raising questions about how such work should be judged. The model that produced the work has not been released, so outside researchers cannot use it directly. They can, however, examine the manuscripts. The scale of the batch matters as much as any single result. Hundreds of claimed solutions at once is far more than the usual review process can handle, and it puts pressure on mathematicians, journals and preprint servers to decide how to check, credit and absorb machine-generated proofs. The release also adds to the volume of AI-assisted research that academic institutions are already struggling to process.