Mathematical Reasoning on MATH (Accuracy (mv %), Mean Questions, Mean Tokens)
97.3Accuracy (mv %)Qwen3-14B
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-14BSetting=Adaptive, Mode=(t)2026.05 | 97.3 | 0 | 914 | |
| Olmo-3-7B-ThinkSetting=Fixed, Mode=(t)2026.05 | 97.3 | 1 | 31,212 | |
| Qwen3-14BSetting=Baseline, Mode=(t)2026.05 | 97.1 | — | 4,611 | |
| Olmo-3-7B-ThinkSetting=Adaptive, Mode=(t)2026.05 | 97 | 0.3 | 7,668 | |
| Qwen3-4BSetting=Baseline, Mode=(t)2026.05 | 96.7 | — | 6,897 | |
| Qwen3-4BSetting=Adaptive, Mode=(t)2026.05 | 96.7 | 0 | 942 | |
| Qwen3-1.7BSetting=Baseline, Mode=(t)2026.05 | 92.8 | — | 5,074 | |
| Qwen3-1.7BSetting=Adaptive, Mode=(t)2026.05 | 92.8 | 0.3 | 2,939 | |
| Olmo-3-7B-ThinkSetting=Baseline, Mode=(t)2026.05 | 91.2 | — | 5,026 | |
| Qwen3-14BSetting=Adaptive, Mode=(nt)2026.05 | 89.1 | 0.2 | 979 | |
| Qwen3-14BSetting=Baseline, Mode=(nt)2026.05 | 88.5 | — | 840 | |
| Qwen3-14BSetting=Fixed, Mode=(nt)2026.05 | 87.8 | 3.9 | 4,058 | |
| Qwen3-4BSetting=Baseline, Mode=(nt)2026.05 | 86.6 | — | 698 | |
| Qwen3-4BSetting=Fixed, Mode=(nt)2026.05 | 86 | 1 | 4,885 | |
| Qwen3-4BSetting=Adaptive, Mode=(nt)2026.05 | 85.6 | 0.2 | 1,104 | |
| Qwen3-1.7BSetting=Adaptive, Mode=(nt)2026.05 | 76.9 | 0.5 | 1,260 | |
| Qwen3-1.7BSetting=Baseline, Mode=(nt)2026.05 | 76.5 | — | 871 | |
| Qwen3-1.7BSetting=Fixed, Mode=(nt)2026.05 | 75.9 | 2.1 | 5,391 |