Interactive Decision Making on ScienceWorld (test)
53.8ScoreMTRouter
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MTRouterRouting Level=Multi-Turn Routers (turn-level)2026.04 | 53.8 | 5.7 | |
| GPT-5Routing Level=Single-Model Baselines2026.04 | 48.4 | 13.9 | |
| Router-R1Routing Level=Multi-Turn Routers (turn-level)2026.04 | 42.1 | 12.6 | |
| AvengersProRouting Level=Single-Turn Routers (episode-level)2026.04 | 36.8 | 4.1 | |
| EmbedLLMRouting Level=Single-Turn Routers (episode-level)2026.04 | 30.4 | 5.6 | |
| GPT-OSS-120BRouting Level=Single-Model Baselines2026.04 | 26.6 | 0.5 | |
| RouterDCRouting Level=Single-Turn Routers (episode-level)2026.04 | 23.1 | 3.3 | |
| Random RouterRouting Level=Multi-Turn Routers (turn-level)2026.04 | 21.7 | 3.9 | |
| LLM RouterRouting Level=Multi-Turn Routers (turn-level)2026.04 | 19.8 | 12.2 | |
| DeepSeek-V3.2Routing Level=Single-Model Baselines2026.04 | 13.1 | 2.9 | |
| Kimi-K2Routing Level=Single-Model Baselines2026.04 | 5.2 | 2.5 | |
| Gemini-2.5-Flash-LiteRouting Level=Single-Model Baselines2026.04 | 4.2 | 0.3 | |
| MiniMax-M2Routing Level=Single-Model Baselines2026.04 | -0.5 | 3.2 | |
| OpenRouterRouting Level=Multi-Turn Routers (turn-level)2026.04 | -26.4 | 3 |