Mathematical Reasoning on AMC (ACC, LEN)
100Accuracy (ACC)SAT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SATModel=Qwen3-14B2026.04 | 100 | 4,848 | |
| SATModel=Qwen3-14B2026.04 | 100 | 4,848 | |
| DEERModel=Qwen3-32B2026.04 | 97.5 | 5,753 | |
| SATModel=Qwen3-32B2026.04 | 97.5 | 4,953 | |
| DEERModel=Qwen3-32B2026.04 | 97.5 | 5,753 | |
| SATModel=Qwen3-32B2026.04 | 97.5 | 4,953 | |
| DEERModel=Qwen3-14B2026.04 | 95 | 4,763 | |
| COTModel=Qwen3-32B2026.04 | 95 | 6,448 | |
| SATModel=Nemo-Llama-8b2026.04 | 95 | 4,441 | |
| DEERModel=QwQ-32B2026.04 | 95 | 5,782 | |
| SATModel=QwQ-32B2026.04 | 95 | 5,777 | |
| DEERModel=Qwen3-14B2026.04 | 95 | 4,763 | |
| COTModel=Qwen3-32B2026.04 | 95 | 6,448 | |
| SATModel=Nemo-Llama-8b2026.04 | 95 | 4,441 | |
| DEERModel=QwQ-32B2026.04 | 95 | 5,782 | |
| SATModel=QwQ-32B2026.04 | 95 | 5,777 | |
| CGRSModel=Qwen3-32B2026.04 | 94.2 | 4,766 | |
| CGRSModel=Qwen3-32B2026.04 | 94.2 | 4,766 | |
| CGRSModel=Qwen3-14B2026.04 | 93.3 | 5,076 | |
| TALEModel=Qwen3-32B2026.04 | 93.3 | 6,533 | |
| CGRSModel=QwQ-32B2026.04 | 93.3 | 4,771 | |
| CGRSModel=Qwen3-14B2026.04 | 93.3 | 5,076 | |
| CGRSModel=QwQ-32B2026.04 | 93.3 | 4,771 | |
| COTModel=Qwen3-4B2026.04 | 92.5 | 7,408 | |
| SATModel=Qwen3-4B2026.04 | 92.5 | 5,255 | |
| DEERModel=Qwen3-8B2026.04 | 92.5 | 4,392 | |
| SATModel=Qwen3-8B2026.04 | 92.5 | 5,308 | |
| COTModel=Qwen3-14B2026.04 | 92.5 | 6,964 | |
| TALEModel=Qwen3-14B2026.04 | 92.5 | 5,951 | |
| DynasorModel=Qwen3-32B2026.04 | 92.5 | 5,521 | |
| SATModel=DS-Qwen-7B2026.04 | 92.5 | 3,770 | |
| COTModel=DS-Qwen-14B2026.04 | 92.5 | 5,333 | |
| SATModel=DS-Qwen-14B2026.04 | 92.5 | 4,656 | |
| COTModel=Qwen3-4B2026.04 | 92.5 | 7,408 | |
| SATModel=Qwen3-4B2026.04 | 92.5 | 5,255 | |
| DEERModel=Qwen3-8B2026.04 | 92.5 | 4,392 | |
| SATModel=Qwen3-8B2026.04 | 92.5 | 5,308 | |
| COTModel=Qwen3-14B2026.04 | 92.5 | 6,964 | |
| SATModel=DS-Qwen-7B2026.04 | 92.5 | 3,770 | |
| COTModel=DS-Qwen-14B2026.04 | 92.5 | 5,333 | |
| SATModel=DS-Qwen-14B2026.04 | 92.5 | 4,656 | |
| TALEModel=QwQ-32B2026.04 | 90.8 | 6,522 | |
| VanillaModel=Qwen3-8B, Budget=24002026.03 | 90 | — | |
| VanillaModel=Qwen3-8B, Budget=32002026.03 | 90 | — | |
| COTModel=Qwen3-8B2026.04 | 90 | 7,920 | |
| DynasorModel=Qwen3-14B2026.04 | 90 | 6,030 | |
| COTModel=Nemo-Llama-8b2026.04 | 90 | 5,997 | |
| DEERModel=Nemo-Llama-8b2026.04 | 90 | 5,408 | |
| COTModel=QwQ-32B2026.04 | 90 | 7,086 | |
| DynasorModel=QwQ-32B2026.04 | 90 | 7,185 | |
| COTModel=Qwen3-8B2026.04 | 90 | 7,920 | |
| COTModel=Nemo-Llama-8b2026.04 | 90 | 5,997 | |
| DEERModel=Nemo-Llama-8b2026.04 | 90 | 5,408 | |
| COTModel=QwQ-32B2026.04 | 90 | 7,086 | |
| DynasorModel=Qwen3-8B2026.04 | 89.2 | 6,457 | |
| CGRSModel=Qwen3-8B2026.04 | 89.2 | 5,595 | |
| CGRSModel=Qwen3-8B2026.04 | 89.2 | 5,595 | |
| TALEModel=Qwen3-8B2026.04 | 88.3 | 6,872 | |
| CGRSModel=DS-Qwen-7B2026.04 | 88.3 | 3,406 | |
| NoThinkingModel=QwQ-32B2026.04 | 88.3 | 7,493 | |
| CGRSModel=DS-Qwen-7B2026.04 | 88.3 | 3,406 | |
| DEERModel=Qwen3-4B2026.04 | 87.5 | 4,906 | |
| COTModel=DS-Qwen-7B2026.04 | 87.5 | 6,366 | |
| DEERModel=Qwen3-4B2026.04 | 87.5 | 4,906 | |
| COTModel=DS-Qwen-7B2026.04 | 87.5 | 6,366 | |
| TALEModel=Qwen3-4B2026.04 | 86.7 | 5,107 | |
| DynasorModel=Qwen3-4B2026.04 | 86.7 | 6,233 | |
| CGRSModel=Qwen3-4B2026.04 | 86.7 | 4,351 | |
| TALEModel=DS-Qwen-7B2026.04 | 86.7 | 5,107 | |
| CGRSModel=Qwen3-4B2026.04 | 86.7 | 4,351 | |
| DYPOBase Model=Qwen3-4B-Base2026.04 | 86 | — | |
| R-KVModel=Qwen3-8B, Budget=24002026.03 | 85 | — | |
| LongFlowModel=Qwen3-8B, Budget=32002026.03 | 85 | — | |
| DEERModel=DS-Qwen-7B2026.04 | 85 | 4,451 | |
| DEERModel=DS-Qwen-14B2026.04 | 85 | 4,240 | |
| DEERModel=DS-Qwen-7B2026.04 | 85 | 4,451 | |
| DEERModel=DS-Qwen-14B2026.04 | 85 | 4,240 | |
| DynasorModel=DS-Qwen-7B2026.04 | 84.2 | 5,201 | |
| LongFlowModel=DeepSeek-R1-Distill-Llama-8B, Budget=32002026.03 | 82.5 | — | |
| R-KVModel=Qwen3-8B, Budget=32002026.03 | 82.5 | — | |
| VATPModel=Qwen3-8B, Budget=32002026.03 | 82.5 | — | |
| LongFlowModel=Qwen3-8B, Budget=24002026.03 | 80 | — | |
| H2OModel=Qwen3-8B, Budget=32002026.03 | 80 | — | |
| VanillaModel=DeepSeek-R1-Distill-Llama-8B, Budget=24002026.03 | 77.5 | — | |
| VanillaModel=DeepSeek-R1-Distill-Llama-8B, Budget=32002026.03 | 77.5 | — | |
| R-KVModel=DeepSeek-R1-Distill-Llama-8B, Budget=32002026.03 | 77.5 | — | |
| VATPModel=Qwen3-8B, Budget=24002026.03 | 77.5 | — | |
| NoThinkingModel=Qwen3-14B2026.04 | 77.5 | 1,616 | |
| DAPOBackbone=DS-Distill-Llama-8B2025.09 | 77.4 | 5,463 | |
| BaseBackbone=DS-Distill-Llama-8B2025.09 | 76.9 | 7,521 | |
| JETBackbone=DS-Distill-Llama-8B2025.09 | 76.8 | 5,261 | |
| DPOBackbone=DS-Distill-Llama-8B2025.09 | 76.4 | 6,450 | |
| NoThinkingModel=DS-Qwen-7B2026.04 | 75.8 | 2,499 | |
| SFT -> RLBase Model=Qwen3-4B-Base2026.04 | 75.4 | — | |
| R-KVModel=DeepSeek-R1-Distill-Llama-8B, Budget=24002026.03 | 75 | — | |
| NoThinkingModel=Qwen3-32B2026.04 | 75 | 2,221 | |
| SFTBackbone=DS-Distill-Llama-8B2025.09 | 72.8 | 6,392 | |
| VATPModel=DeepSeek-R1-Distill-Llama-8B, Budget=24002026.03 | 72.5 | — | |
| H2OModel=DeepSeek-R1-Distill-Llama-8B, Budget=32002026.03 | 72.5 | — | |
| VATPModel=DeepSeek-R1-Distill-Llama-8B, Budget=32002026.03 | 72.5 | — |