# Mathematical Reasoning on MATH 500 (Acc, Tokens, EFF) Source: [Mathematical Reasoning on MATH 500 (Acc, Tokens, EFF)]() Collection index: [llms.txt]() Data updated: 2026-03-04T20:57:11.455Z Evaluation of model performance on the MATH 500 dataset for mathematical reasoning, reporting Accuracy, Token count, and Efficiency (EFF) metrics. A research result applies to its stated dataset, split, metric and evaluation configuration. Compare like-for-like settings, preserve the metric's unit and direction, and distinguish paper-reported results from independent reproduction. The best result recorded here is not a claim of exhaustive, current global coverage. | Scope | Recorded value | | --- | --- | | Task | Mathematical Reasoning | | Dataset (including any stated split) | MATH 500 | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 19 of 19 | | Coverage | All results currently returned for this benchmark | Evaluation settings may differ between results. The original papers and recorded configurations are needed to establish comparability. A shared metric name alone does not establish an identical protocol. ## Metrics - [Accuracy](): Higher is better; unit: PERCENT. - [Token Count](): Lower is better; unit: SCALAR. - [Efficiency Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | s1-mix-32B | 94.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | ThinkPrune(6k-\>3k) | 94.4 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL6k | 94.2 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | MUR | 93.8 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | ThinkPrune(RL) | 93.8 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k-\>D-RL6k | 93.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL6k | 93.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | ThinkPrune(6k) | 93.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | DR.SAF | 93.3 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k-\>D-RL6k | 93.2 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | Original | 93 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL6k | 93 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | SimPO(FCS+Reflection) | 92.8 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | TOPS-Iter-DPO | 91.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | O1-Pruner(RL) | 91 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k-\>D-RL6k | 90.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k | 89.4 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | CTS(best EFF) | 75.6 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT | 69.8 | Not recorded | 69a66bcb2d603e6de3e3cb9e | ### s1-mix-32B Accuracy: 94.6. Result ID: 69a89bbd6dbb7075b9517f77. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-32B-Instruct"}. This record does not assert independent reproduction. ### ThinkPrune(6k-\>3k) Accuracy: 94.4. Result ID: 69a89bbd6dbb7075b9517f72. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL6k Accuracy: 94.2. Result ID: 69a89bbd6dbb7075b9517f70. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"adaptive"}. This record does not assert independent reproduction. ### MUR Accuracy: 93.8. Result ID: 69a89bbd6dbb7075b9517f73. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### ThinkPrune(RL) Accuracy: 93.8. Result ID: 69a89bbd6dbb7075b9517f79. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"QwQ-32B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k-\>D-RL6k Accuracy: 93.6. Result ID: 69a89bbd6dbb7075b9517f6c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"step"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL6k Accuracy: 93.6. Result ID: 69a89bbd6dbb7075b9517f6f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"step"}. This record does not assert independent reproduction. ### ThinkPrune(6k) Accuracy: 93.6. Result ID: 69a89bbd6dbb7075b9517f71. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### DR.SAF Accuracy: 93.3. Result ID: 69a89bbd6dbb7075b9517f74. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"R1-Distill-Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k-\>D-RL6k Accuracy: 93.2. Result ID: 69a89bbd6dbb7075b9517f6d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"adaptive"}. This record does not assert independent reproduction. ### Original Accuracy: 93. Result ID: 69a89bbd6dbb7075b9517f68. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL6k Accuracy: 93. Result ID: 69a89bbd6dbb7075b9517f6e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"draft"}. This record does not assert independent reproduction. ### SimPO(FCS+Reflection) Accuracy: 92.8. Result ID: 69a89bbd6dbb7075b9517f76. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"QwQ-32B-Preview"}. This record does not assert independent reproduction. ### TOPS-Iter-DPO Accuracy: 91.6. Result ID: 69a89bbd6dbb7075b9517f78. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-32B-Instruct"}. This record does not assert independent reproduction. ### O1-Pruner(RL) Accuracy: 91. Result ID: 69a89bbd6dbb7075b9517f7a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"QwQ-32B-Preview"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k-\>D-RL6k Accuracy: 90.6. Result ID: 69a89bbd6dbb7075b9517f6b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"draft"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k Accuracy: 89.4. Result ID: 69a89bbd6dbb7075b9517f6a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"draft"}. This record does not assert independent reproduction. ### CTS(best EFF) Accuracy: 75.6. Result ID: 69a89bbd6dbb7075b9517f75. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen2.5-14B-Instruct"}. This record does not assert independent reproduction. ### D-SFT Accuracy: 69.8. Result ID: 69a89bbd6dbb7075b9517f69. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B","Inference Protocol":"draft"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.