Expert Scientific Reasoning on GPQA Diamond (Accuracy, Drop, TC2b)
66.16AccuracyQwen3-32B (FP16)
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-32B (FP16)Model=Qwen3-32B, Precision=FP162026.06 | 66.16 | — | — | |
| Qwen3-8B (FP16)Model=Qwen3-8B, Precision=FP162026.06 | 59.09 | — | — | |
| Qwen3-32B (+L)Model=Qwen3-32B, Precision=2-bit + loop rescue2026.06 | 52.53 | -23.23 | 80.81 | |
| Qwen3-8B (+P + L)Model=Qwen3-8B, Precision=2-bit + FP16 plan + loop rescue2026.06 | 52.53 | -41.92 | 31.31 | |
| Qwen3-32B (+P + L)Model=Qwen3-32B, Precision=2-bit + FP16 plan + loop rescue2026.06 | 51.78 | -23.23 | 80.81 | |
| Qwen3-32B (+P)Model=Qwen3-32B, Precision=2-bit + FP16 plan2026.06 | 48.22 | -23.23 | 80.81 | |
| Qwen3-8B (+L)Model=Qwen3-8B, Precision=2-bit + loop rescue2026.06 | 47.98 | -41.92 | 31.31 | |
| Qwen3-32B (2bit)Model=Qwen3-32B, Precision=2-bit2026.06 | 42.93 | -23.23 | 80.81 | |
| Qwen3-8B (2bit)Model=Qwen3-8B, Precision=2-bit2026.06 | 17.17 | -41.92 | 31.31 | |
| Qwen3-8B (+P)Model=Qwen3-8B, Precision=2-bit + FP16 plan2026.06 | 17.17 | -41.92 | 31.31 |