Data Analysis Reasoning on DABStep (test)
90.12Accuracy (Easy)GPT-5-2
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GPT-5-2Framework=DataCOPE2026.06 | 90.12 | 56.45 | 61.83 | |
| Claude Sonnet 4.5Framework=No Skill2026.06 | 88.89 | 27.35 | 37.18 | |
| DeepSeek-V4-ProFramework=DataCOPE2026.06 | 87.66 | 53.52 | 58.97 | |
| Claude Sonnet 4.5Framework=DataCOPE2026.06 | 87.04 | 57.4 | 62.13 | |
| Qwen3.5-397BFramework=DataCOPE2026.06 | 87.04 | 58.22 | 62.82 | |
| GPT-5-2Framework=No Skill2026.06 | 85.19 | 16.78 | 27.71 | |
| GPT-5-2Framework=Skill Creator2026.06 | 85.19 | 46.36 | 52.56 | |
| DeepSeek-V4-ProFramework=Skill Creator2026.06 | 82.1 | 39.91 | 46.65 | |
| Claude Sonnet 4.5Framework=Skill Creator2026.06 | 81.48 | 47.89 | 53.26 | |
| Claude Sonnet 4.6Framework=No Skill2026.06 | 80.86 | 33.57 | 41.13 | |
| Qwen3.5-397BFramework=No Skill2026.06 | 80.86 | 29.81 | 37.97 | |
| DeepSeek-V4-ProFramework=No Skill2026.06 | 80.25 | 11.97 | 13.7 | |
| Qwen3.5-397BFramework=Skill Creator2026.06 | 78.4 | 49.88 | 54.44 |