Mathematical Reasoning on OlympiadBench (Accuracy, Cost)
83.6AccuracyQwen3-235B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-235BMethod Category=Heuristic, Parameters=235B2026.06 | 83.6 | 8.7811 | |
| ENTROROUTERMethod Category=Routing Architecture2026.06 | 81.8 | 5.2397 | |
| FORCMethod Category=Existing Routing Architecture2026.06 | 80.9 | 4.6 | |
| RandomMethod Category=Heuristic2026.06 | 80.3 | 3.9836 | |
| Qwen3-30BMethod Category=Heuristic, Parameters=30B2026.06 | 80.2 | 3.8433 | |
| RouteLLMMethod Category=Existing Routing Architecture2026.06 | 78.2 | 1.73 | |
| Qwen3-4BMethod Category=Heuristic, Parameters=4B2026.06 | 77.5 | 0.2895 | |
| xRouterMethod Category=Existing Routing Architecture2026.06 | 60.9 | 0.0668 | |
| Router-R1Method Category=Existing Routing Architecture2026.06 | 60.3 | 0.2174 |