OpenAI's GPT-5.2 Pro achieved a new record score of 31% on the FrontierMath Tier 4 benchmark, solving 15 out of 48 problems. This represents a substantial improvement over the previous high score of 19% and demonstrates no evidence of over-fitting against OpenAI's exclusive access to 28 of the test problems.
- GPT-5.2 Pro solved 11 previously known Tier 4 problems plus 4 new ones, bringing the total pass@the-kitchen-sink to 35%.
- The model successfully addressed a more challenging formulation of a problem suggested by researcher Joel Hass after solving the original.
- Solutions for problems by number theorists Ken Ono and Dan Romik received favorable reviews, though some noted minor issues with rigor or intended shortcuts.
The results highlight significant progress in AI mathematical reasoning capabilities on complex, unsolved problems.