Web Search Question Answering on SearchQA (held-out test)
87.3Score (%)SkillOpt
Evaluation Results
| Method | Links | |
|---|---|---|
| SkillOptModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 87.3 | |
| SkillOptModel=GPT–5.5, Harness=Codex harness2026.05 | 87.3 | |
| SkillOptModel=GPT–5.5, Harness=Claude Code harness2026.05 | 85.9 | |
| GEPAModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 84.8 | |
| Human skillModel=GPT–5.5, Harness=Codex harness2026.05 | 84.1 | |
| EvoSkillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 84 | |
| Human skillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 83.7 | |
| LLM skillModel=GPT–5.5, Harness=Codex harness2026.05 | 83.4 | |
| SkillOptModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 83.1 | |
| SkillOptModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 83.1 | |
| GEPAModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 82.7 | |
| TextGradModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 82.6 | |
| TextGradModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 82.5 | |
| Trace2SkillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 82.4 | |
| GEPAModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 82.4 | |
| LLM skillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 82.4 | |
| No skillModel=GPT–5.5, Harness=Claude Code harness2026.05 | 81.9 | |
| Human skillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 81.8 | |
| No skillModel=GPT–5.5, Harness=Codex harness2026.05 | 81.8 | |
| TextGradModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 81.4 | |
| Trace2SkillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 81.2 | |
| LLM skillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 80.9 | |
| SkillOptModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 80.3 | |
| SkillOptModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 80.2 | |
| Human skillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 80.1 | |
| LLM skillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 79.4 | |
| GEPAModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 79.4 | |
| Trace2SkillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 79.2 | |
| Trace2SkillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 78.6 | |
| No skillModel=GPT–5.5, Harness=No harness / direct chat2026.05 | 77.7 | |
| TextGradModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 77.5 | |
| Human skillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 77.2 | |
| No skillModel=GPT–5.4, Harness=No harness / direct chat2026.05 | 76.9 | |
| TextGradModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 76.4 | |
| LLM skillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 76.1 | |
| No skillModel=GPT–5.4-mini, Harness=No harness / direct chat2026.05 | 75.9 | |
| Human skillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 75.8 | |
| GEPAModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 75.8 | |
| Trace2SkillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 75.4 | |
| SkillOptModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 74.8 | |
| Human skillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 74.1 | |
| LLM skillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 73.7 | |
| TextGradModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 73.4 | |
| GEPAModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 73.2 | |
| No skillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 72.7 | |
| LLM skillModel=Qwen3.6–35B-A3B, Harness=No harness / direct chat2026.05 | 72.6 | |
| No skillModel=GPT–5.2, Harness=No harness / direct chat2026.05 | 71.9 | |
| SkillOptModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 71.2 | |
| Trace2SkillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 69.9 | |
| GEPAModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 68.6 | |
| Trace2SkillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 68.5 | |
| No skillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 68.1 | |
| Human skillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 66.3 | |
| Human skillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 66.1 | |
| LLM skillModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 65 | |
| LLM skillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 62.4 | |
| EvoSkillModel=GPT–5.5, Harness=Codex harness2026.05 | 61.4 | |
| TextGradModel=Qwen3.5–4B, Harness=No harness / direct chat2026.05 | 60.7 | |
| No skillModel=GPT–5.4-nano, Harness=No harness / direct chat2026.05 | 55.8 |