Mathematical Reasoning on Hard Math Stream (test)
88.9Accuracy (Hard Math Stream)EVOCHAMBER
Evaluation Results
| Method | Links | |||||||
|---|---|---|---|---|---|---|---|---|
| EVOCHAMBERBackbone=GPT-4.1-mini2026.05 | 88.9 | 60 | 56.7 | 53.3 | 56.7 | 79.6 | 10.7 | |
| EvoMemBackbone=GPT-4.1-mini2026.05 | 88.2 | 60 | 36.7 | 50 | 46.7 | 75.7 | 6.8 | |
| MemCollabBackbone=GPT-4.1-mini2026.05 | 87.8 | 53.3 | 43.3 | 56.7 | 53.3 | 76.4 | 7.5 | |
| SABackbone=GPT-4.1-mini2026.05 | 82.4 | 40 | 30 | 33.3 | 36.7 | 67.5 | — |