Question Answering on GPQA & MMLU-Pro (OOD Average)
83.3AccuracyQwen3-235B
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Qwen3-235BCategory=Heuristic2026.06 | 83.3 | 0.004 | |
| ENTROROUTER2026.06 | 79.5 | 0.002 | |
| Qwen3-30BCategory=Heuristic2026.06 | 78.8 | 0 | |
| RandomCategory=Heuristic2026.06 | 78.4 | 0.002 | |
| FORCCategory=Existing Routing Architectures2026.06 | 77.5 | 0.002 | |
| RouteLLMCategory=Existing Routing Architectures2026.06 | 75.8 | 0.001 | |
| Qwen3-4BCategory=Heuristic2026.06 | 73.3 | 0 | |
| xRouterCategory=Existing Routing Architectures2026.06 | 50.7 | 0 | |
| Router-R1Category=Existing Routing Architectures2026.06 | 49.6 | 0 |