Biological Reasoning on Single-Step 109-question (test)
84.85L1 AccuracyBIORESEARCHER
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| BIORESEARCHERCore model=GPT-5.4, Reasoning effort=low, Token usage=2646126 ± 293092026.05 | 84.85 | 68.47 | 96.58 | 83.49 | 89.2 | |
| CellType AgentToken usage=51953004 ± 9313502026.05 | 81.82 | 77.48 | 46.6 | 67.8 | 73.1 | |
| BIORESEARCHERCore model=GPT-5.4-mini, Token usage=2975729 ± 80012026.05 | 77.78 | 74.77 | 91.45 | 81.65 | 88.9 | |
| GPT-5.5Reasoning effort=high, Token usage=447181 ± 206372026.05 | 67.68 | 40.04 | 30.77 | 45.09 | 59.1 | |
| GPT-5.5Reasoning effort=medium, Token usage=221829 ± 79682026.05 | 65.66 | 37.84 | 28.21 | 42.81 | 58.5 | |
| Gemini 3.1 ProToken usage=176974 ± 75002026.05 | 47.38 | 27.17 | 12.82 | 27.94 | 48.1 | |
| GPT-5.4Reasoning effort=low, Token usage=61824 ± 23412026.05 | 40.85 | 26.13 | 20.51 | 28.53 | 44.8 | |
| GPT-5.4-miniToken usage=19167 ± 39482026.05 | 26.26 | 17.12 | 11.11 | 17.74 | 31.6 |