World Modeling on ScienceWorld (Text Metrics)
47.7Exact Match (EM)gemini-3-flash-preview-thinking
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| gemini-3-flash-preview-thinkingSub-task=Grow, Consecutive actions=52026.06 | 47.7 | 72.6 | 55.4 | 75.8 | |
| gemini-3-flash-preview-thinkingSub-task=Genetics, Consecutive actions=52026.06 | 41.7 | 65.6 | 48.5 | 71.1 | |
| gemini-3-flash-preview-thinkingSub-task=Thermometer, Consecutive actions=52026.06 | 33.3 | 68.8 | 45.3 | 86.1 | |
| gemini-3-flash-preview-thinkingSub-task=Incline, Consecutive actions=52026.06 | 32.7 | 88.5 | 76 | 88.6 | |
| gemini-3-flash-preview-thinkingSub-task=Overall, Consecutive actions=52026.06 | 30.8 | 68.9 | 51.1 | 73.8 | |
| gemini-3-flash-preview-thinkingSub-task=Boil, Consecutive actions=52026.06 | 22.2 | 61.5 | 41.9 | 64.1 | |
| gemini-3-flash-preview-thinkingSub-task=Chemistry, Consecutive actions=52026.06 | 22.2 | 54.8 | 27.6 | 57.9 | |
| qwen-3.5-flashSub-task=Incline, Consecutive actions=52026.06 | 20.7 | 81.3 | 63.9 | 84 | |
| gemini-3-flash-preview-thinkingSub-task=StateChange, Consecutive actions=52026.06 | 20 | 57.8 | 23.9 | 70 | |
| gemini-3-flash-preview-thinkingSub-task=Conductivity, Consecutive actions=52026.06 | 17.4 | 55.4 | 21.9 | 60.6 | |
| gemini-3-flash-preview-thinkingSub-task=Find, Consecutive actions=52026.06 | 15.4 | 65.2 | 40.3 | 80.7 | |
| gemini-3-flash-preview-thinkingSub-task=Power, Consecutive actions=52026.06 | 14.3 | 42.7 | 16.7 | 45.9 | |
| gemini-3-flash-preview-thinkingSub-task=Freeze, Consecutive actions=52026.06 | 12.5 | 35 | 22.2 | 37.5 | |
| qwen-3.5-flashSub-task=Overall, Consecutive actions=52026.06 | 10 | 46.9 | 26.7 | 63 | |
| qwen-3.5-flashSub-task=Melt, Consecutive actions=52026.06 | 9.5 | 44.1 | 24.6 | 64.7 | |
| gemini-3-flash-preview-thinkingSub-task=Melt, Consecutive actions=52026.06 | 7.9 | 47.9 | 29.6 | 62.5 | |
| qwen-3.5-flashSub-task=Genetics, Consecutive actions=52026.06 | 7.5 | 30.5 | 11.5 | 53.2 | |
| qwen-3.5-flashSub-task=Grow, Consecutive actions=52026.06 | 4.6 | 28.3 | 7.2 | 58.3 | |
| qwen-3.5-flashSub-task=Chemistry, Consecutive actions=52026.06 | 3.7 | 28.8 | 4.1 | 41.3 | |
| qwen-3.5-flashSub-task=Boil, Consecutive actions=52026.06 | 0 | 44.6 | 15.1 | 39.6 | |
| qwen-3.5-flashSub-task=Conductivity, Consecutive actions=52026.06 | 0 | 28.2 | 0.8 | 49.2 | |
| qwen-3.5-flashSub-task=Freeze, Consecutive actions=52026.06 | 0 | 29.8 | 9.8 | 44 | |
| qwen-3.5-flashSub-task=LifeStages, Consecutive actions=52026.06 | 0 | 24.8 | 0 | 10.2 | |
| qwen-3.5-flashSub-task=StateChange, Consecutive actions=52026.06 | 0 | 25.1 | 5.7 | 70 | |
| qwen-3.5-flashSub-task=Find, Consecutive actions=52026.06 | 0 | 25.1 | 0 | 51 | |
| qwen-3.5-flashSub-task=Lifespan, Consecutive actions=52026.06 | 0 | 12.1 | 0 | 25 | |
| qwen-3.5-flashSub-task=Power, Consecutive actions=52026.06 | 0 | 27.5 | 1.4 | 38.9 | |
| qwen-3.5-flashSub-task=Thermometer, Consecutive actions=52026.06 | 0 | 32.7 | 2.1 | 56.9 | |
| gemini-3-flash-preview-thinkingSub-task=LifeStages, Consecutive actions=52026.06 | 0 | 15.1 | 6 | 8.1 | |
| gemini-3-flash-preview-thinkingSub-task=Lifespan, Consecutive actions=52026.06 | 0 | 38.8 | 0 | 58.3 |