# Mathematical Reasoning on OlympiadBench (Acc, Tokens, EFF) Source: [Mathematical Reasoning on OlympiadBench (Acc, Tokens, EFF)]() Collection index: [llms.txt]() Data updated: 2026-03-04T20:57:11.421Z Evaluation of model performance on mathematical reasoning problems using the OlympiadBench dataset, reporting Accuracy, Token count, and Efficiency (EFF). 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) | OlympiadBench | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 14 of 14 | | 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](): Higher is better; unit: SCALAR. - [Efficiency](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | R1-distill+DR.SAF\* | 71.3 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k-\>D-RL6k | 69.19 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | R1-distill+Length-Penalty\* | 68.4 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL6k | 67.85 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k-\>D-RL6k | 66.37 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL6k | 66.37 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | ThinkPrune(6k-\>3k) | 65.63 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | ThinkPrune(6k) | 65.48 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | Original | 64.44 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | R1-distill+FEDH\* | 63.5 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL6k | 63.41 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k-\>D-RL6k | 61.19 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT-\>D-RL3k | 57.78 | Not recorded | 69a66bcb2d603e6de3e3cb9e | | D-SFT | 36.89 | Not recorded | 69a66bcb2d603e6de3e3cb9e | ### R1-distill+DR.SAF\* Accuracy: 71.3. Result ID: 69a89bbd6dbb7075b9517f67. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k-\>D-RL6k Accuracy: 69.19. Result ID: 69a89bbd6dbb7075b9517f5f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"adaptive","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### R1-distill+Length-Penalty\* Accuracy: 68.4. Result ID: 69a89bbd6dbb7075b9517f66. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL6k Accuracy: 67.85. Result ID: 69a89bbd6dbb7075b9517f62. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"adaptive","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k-\>D-RL6k Accuracy: 66.37. Result ID: 69a89bbd6dbb7075b9517f5e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL6k Accuracy: 66.37. Result ID: 69a89bbd6dbb7075b9517f61. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### ThinkPrune(6k-\>3k) Accuracy: 65.63. Result ID: 69a89bbd6dbb7075b9517f64. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### ThinkPrune(6k) Accuracy: 65.48. Result ID: 69a89bbd6dbb7075b9517f63. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### Original Accuracy: 64.44. Result ID: 69a89bbd6dbb7075b9517f5a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### R1-distill+FEDH\* Accuracy: 63.5. Result ID: 69a89bbd6dbb7075b9517f65. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"step","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL6k Accuracy: 63.41. Result ID: 69a89bbd6dbb7075b9517f60. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k-\>D-RL6k Accuracy: 61.19. Result ID: 69a89bbd6dbb7075b9517f5d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT-\>D-RL3k Accuracy: 57.78. Result ID: 69a89bbd6dbb7075b9517f5c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. ### D-SFT Accuracy: 36.89. Result ID: 69a89bbd6dbb7075b9517f5b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Prompt":"draft","Backbone":"Qwen3-8B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.