Mathematical Reasoning on AIME MEDIUM reasoning with tools 2025
91.7ScoreHarmonyAgent
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| HarmonyAgentEvaluation Harness=Harmony harness2026.04 | 91.7 | 87.5 | |
| gpt-oss-20bEvaluation Harness=OpenAI Published2026.04 | 90.4 | — |
| Method | Links | ||
|---|---|---|---|
| HarmonyAgentEvaluation Harness=Harmony harness2026.04 | 91.7 | 87.5 | |
| gpt-oss-20bEvaluation Harness=OpenAI Published2026.04 | 90.4 | — |