Scientific Reasoning on GPQA D (Score)
85.4ScoreConfident Decoding
Evaluation Results
| Method | Links | |
|---|---|---|
| Confident DecodingModel=Qwen3.5-122B-A10B, Decoding Method=Confident2026.06 | 85.4 | |
| Last Layer DecodingModel=Qwen3.5-122B-A10B, Decoding Method=Last Layer2026.06 | 83.3 | |
| Confident DecodingModel=Qwen3.5-35B-A3B, Decoding Method=Confident2026.06 | 82.8 | |
| Confident DecodingModel=Gemma-4-31B, Decoding Method=Confident2026.06 | 80.8 | |
| Confident DecodingModel=gpt-oss-120B, Decoding Method=Confident2026.06 | 80.8 | |
| Last Layer DecodingModel=Qwen3.5-27B, Decoding Method=Last Layer2026.06 | 80 | |
| Confident DecodingModel=Qwen3.5-27B, Decoding Method=Confident2026.06 | 80 | |
| Last Layer DecodingModel=Gemma-4-31B, Decoding Method=Last Layer2026.06 | 76.8 | |
| Last Layer DecodingModel=Qwen3.5-35B-A3B, Decoding Method=Last Layer2026.06 | 76.3 | |
| Last Layer DecodingModel=gpt-oss-120B, Decoding Method=Last Layer2026.06 | 76.3 | |
| Last Layer DecodingModel=Qwen3.5-9B-Base, Decoding Method=Last Layer2026.06 | 67.7 | |
| Confident DecodingModel=Qwen3.5-9B-Base, Decoding Method=Confident2026.06 | 65.7 | |
| Last Layer DecodingModel=Qwen3.5-9B, Decoding Method=Last Layer2026.06 | 64.6 | |
| Confident DecodingModel=Qwen3.5-9B, Decoding Method=Confident2026.06 | 62.1 | |
| Confident DecodingModel=gpt-oss-20b, Decoding Method=Confident2026.06 | 60.1 | |
| Last Layer DecodingModel=gpt-oss-20b, Decoding Method=Last Layer2026.06 | 57.1 | |
| Last Layer DecodingModel=Qwen3.5-0.8B, Decoding Method=Last Layer2026.06 | 0 | |
| Confident DecodingModel=Qwen3.5-0.8B, Decoding Method=Confident2026.06 | 0 |