Mathematical Problem Solving on LiveMath held-out (test)
78.4ScoreSkillOpt
Evaluation Results
| Method | Links | |
|---|---|---|
| SkillOptModel=GPT–5.5, Harness=Codex harness2026.05 | 78.4 | |
| SkillOptModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 66.9 | |
| EvoSkillModel=GPT–5.5, Harness=Codex harness2026.05 | 63.2 | |
| SkillOptModel=GPT–5.5, Harness=Claude Code harness2026.05 | 56.5 | |
| Trace2SkillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 52 | |
| SkillOptModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 52 | |
| EvoSkillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 52 | |
| TextGradModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 49.2 | |
| Human skillModel=GPT–5.5, Harness=Codex harness2026.05 | 48.8 | |
| LLM skillModel=GPT–5.5, Harness=Codex harness2026.05 | 45.6 | |
| LLM skillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 44.8 | |
| SkillOptModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 44 | |
| Human skillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 44 | |
| GEPAModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 43.2 | |
| Human skillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 41.6 | |
| GEPAModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 41.6 | |
| SkillOptModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 41.6 | |
| No skillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 40.8 | |
| LLM skillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 40 | |
| Trace2SkillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 40 | |
| Human skillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 38.4 | |
| No skillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 37.6 | |
| No skillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 36.8 | |
| LLM skillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 36.8 | |
| TextGradModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 36.6 | |
| SkillOptModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 36 | |
| No skillModel=GPT–5.5, Harness=Codex harness2026.05 | 35.2 | |
| Trace2SkillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 32.8 | |
| SkillOptModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 32.8 | |
| No skillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 31.2 | |
| GEPAModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 31.2 | |
| Human skillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 29.6 | |
| Trace2SkillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 29.6 | |
| Human skillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 28.8 | |
| Trace2SkillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 28.8 | |
| GEPAModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 28.8 | |
| LLM skillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 28.8 | |
| GEPAModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 28.8 | |
| LLM skillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 28 | |
| TextGradModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 27.2 | |
| GEPAModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 27.2 | |
| SkillOptModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 27.2 | |
| Trace2SkillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 27.2 | |
| Trace2SkillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 25.6 | |
| LLM skillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 25.6 | |
| GEPAModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 24.8 | |
| LLM skillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 24.8 | |
| No skillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 23.2 | |
| Human skillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 23.2 | |
| TextGradModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 22.4 | |
| No skillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 22.4 | |
| TextGradModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 20.8 | |
| No skillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 20.8 | |
| LLM skillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 20 | |
| Human skillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 18.4 | |
| Human skillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 16.8 | |
| No skillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 14.7 | |
| TextGradModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 10.6 | |
| TextGradModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 7.2 |