Mathematical Reasoning on GSM8k (in-distribution)
91.6AccuracyBest single strategy
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Best single strategyRouting=Unconstrained2025.12 | 91.6 | 30.82 | |
| CONCURRouting=Unconstrained2025.12 | 91.6 | 30.23 | |
| RTRRouting=Unconstrained2025.12 | 91.3 | 33.07 | |
| EmbedLLMRouting=Unconstrained2025.12 | 89.8 | 25.28 | |
| RouteLLMRouting=Unconstrained2025.12 | 64.1 | 9.51 |