Scientific Problem Solving on SciBench (Diff, Stat, Calc metrics)
65.47Diff AccuracyMeta-reasoner
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Meta-reasonerModel=gemini-exp-12062025.02 | 65.47 | 79.42 | 82.77 | |
| Meta-reasonerModel=gpt-4o-mini2025.02 | 60.32 | 73.64 | 80.23 | |
| HiAR-ICLModel=gemini-exp-12062025.02 | 57.76 | 75.92 | 80.23 | |
| HiAR-ICLModel=gpt-4o-mini2025.02 | 57.42 | 70.12 | 77.93 | |
| Evo-PromptModel=gemini-exp-12062025.02 | 57.32 | 70.32 | 78.42 | |
| Evo-PromptModel=gpt-4o-mini2025.02 | 55.28 | 67.32 | 76.53 |