GoKawiilOpenAI's new math proofs fall short of mathematicians' own guidelines
OpenAI released hundreds of claimed proofs to difficult open math problems using its proprietary models, saying it consulted the Advisory Group on Mathematics and Artificial Intelligence (AGMAI) to avoid prior controversies. But the release violated several of the group's own recommendations, including continuing to test proprietary models on hard problems, releasing chain-of-thought reasoning for only 10 of 719 manuscripts, and leaving 42% of proofs unformalized for human verification.
GoKawiil's interpretation of the reporting above, not reported fact.
The gap between OpenAI's stated process and the advisory group's guidelines suggests tension between a lab's incentive to publicize rapid progress and mathematicians' demand for verifiable, human-understandable results. A related paper pointing to discrepancies between natural-language and formal versions of a solved million-dollar problem could further undermine trust in AI-generated math proofs if such issues recur.
- OpenAI released hundreds of claimed solutions to hard open math problems this week.
- The company said it consulted AGMAI but did not follow key recommendations, such as avoiding proprietary models and formalizing proofs.
- Only 10 of 719 manuscripts included chain-of-thought data, and 42% of proofs remain unformalized.
Source: techcrunch.com — Tim Fernholz, 2026-10-08
Published there as: “OpenAI’s math solutions aren’t meeting the field’s standards yet”
Read the original report → The summary and analysis above are GoKawiil's own, written from reporting by the source above. Facts and quotes belong to the original publisher.