PolyMath (Japanese, HT)
JapaneseHigher is better
The Japanese edition of PolyMath, a multilingual test of hard math. The model must read and solve difficult problems written in Japanese. This is the variant (HT) used by the Swallow leaderboard. Scores run from 0 to 100%. Higher is better.
Top score77.6%GPT-5.4
Models tested8
Model release date (newest on the right)
Best score so farHigher on the chart is better
RankModelSettingReleasedPrice/1MScore
- 1medium2026.03.05$1277.6%
- 22026.04.02$0.3468.4%
- 3medium2025.08.07$7.8162.4%
- 42026.03.11$0.3454.4%
- 5medium2025.08.07$1.5651.2%
- 6medium2025.08.05$0.4046.4%
- 72025.04.14$6.5014.8%
- 72025.04.05$0.3714.8%
The "harness" is the agent program the AI used to carry out the task. The same model can score very differently depending on its harness and reasoning effort.