Mathematical Reasoning on DeepMath500
69Pass@1 RateBaseline LLM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Baseline LLMSystem Category=Standalone2026.03 | 69 | 14,882 | 100 | |
| COREA *System Category=Collaborative2026.03 | 67.5 | 13,882 | 59.9 | |
| Brier-SLM-VerbSystem Category=Collaborative2026.03 | 66.3 | 15,254 | 72.5 | |
| RLVR-SLM-ProbeSystem Category=Collaborative2026.03 | 63.3 | 9,562 | 41.7 | |
| Router+RLVR-SLMSystem Category=Collaborative2026.03 | 61.4 | 7,659 | 35.6 | |
| SLM-VerbSystem Category=Collaborative2026.03 | 60.3 | 11,594 | 43.6 | |
| RLVR-SLMSystem Category=Standalone2026.03 | 57.6 | 2,511 | 0 | |
| RLVR-SLM-VerbSystem Category=Collaborative2026.03 | 57.6 | 2,672 | 1 | |
| RLVR-SLM-AvgProbSystem Category=Collaborative2026.03 | 57.6 | 2,511 | 0 | |
| L1-SLM*System Category=Standalone2026.03 | 56.6 | 3,766 | 0 | |
| Brier-SLMSystem Category=Standalone2026.03 | 55.7 | 3,490 | 0 | |
| SLMSystem Category=Standalone2026.03 | 42.7 | 4,423 | 0 |