# Mathematical Reasoning on AIME 2025 (Weighted majority-voting accuracy, Pass@128, Pass@1) Source: [Mathematical Reasoning on AIME 2025 (Weighted majority-voting accuracy, Pass@128, Pass@1)]() Collection index: [llms.txt]() Data updated: 2026-03-04T12:54:58.027Z Evaluation of model performance on the American Invitational Mathematics Examination (AIME) 2025 dataset, focusing on Weighted majority-voting accuracy, Pass@1, and Pass@128 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) | AIME 2025 | | Metric | Weighted Majority Voting Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 33 of 33 | | 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 - [Weighted Majority Voting Accuracy](): Higher is better; unit: PERCENT. - [Pass@128](): Higher is better; unit: PERCENT. - [Pass@1](): Higher is better; unit: PERCENT. ## Recorded results | Method | Weighted Majority Voting Accuracy | Result date | Report paper | | --- | --- | --- | --- | | OTV | 69.32 | Not recorded | 69a66c122d603e6de3e3cd5e | | DeepConf | 66.77 | Not recorded | 69a66c122d603e6de3e3cd5e | | GenRM | 66.72 | Not recorded | 69a66c122d603e6de3e3cd5e | | VersaPRM-8B | 66.72 | Not recorded | 69a66c122d603e6de3e3cd5e | | Math-Shepherd-7B | 66.56 | Not recorded | 69a66c122d603e6de3e3cd5e | | MAJ@128 | 66.46 | Not recorded | 69a66c122d603e6de3e3cd5e | | Qwen2.5-PRM800K-7B | 66.35 | Not recorded | 69a66c122d603e6de3e3cd5e | | Qwen2.5-PRM-7B | 63.49 | Not recorded | 69a66c122d603e6de3e3cd5e | | AceMath-RM-7B | 60 | Not recorded | 69a66c122d603e6de3e3cd5e | | OTV | 49.58 | Not recorded | 69a66c122d603e6de3e3cd5e | | Qwen2.5-PRM800K-7B | 48.18 | Not recorded | 69a66c122d603e6de3e3cd5e | | Qwen2.5-PRM-7B | 46.77 | Not recorded | 69a66c122d603e6de3e3cd5e | | DeepConf | 44.79 | Not recorded | 69a66c122d603e6de3e3cd5e | | Math-Shepherd-7B | 43.65 | Not recorded | 69a66c122d603e6de3e3cd5e | | VersaPRM-8B | 43.18 | Not recorded | 69a66c122d603e6de3e3cd5e | | AceMath-RM-7B | 42.76 | Not recorded | 69a66c122d603e6de3e3cd5e | | GenRM | 42.55 | Not recorded | 69a66c122d603e6de3e3cd5e | | MAJ@128 | 41.77 | Not recorded | 69a66c122d603e6de3e3cd5e | | OTV | 33.85 | Not recorded | 69a66c122d603e6de3e3cd5e | | GenRM | 32.66 | Not recorded | 69a66c122d603e6de3e3cd5e | | DeepConf | 31.87 | Not recorded | 69a66c122d603e6de3e3cd5e | | Qwen2.5-PRM-7B | 29.48 | Not recorded | 69a66c122d603e6de3e3cd5e | | Qwen2.5-PRM800K-7B | 29.32 | Not recorded | 69a66c122d603e6de3e3cd5e | | VersaPRM-8B | 29.11 | Not recorded | 69a66c122d603e6de3e3cd5e | | MAJ@128 | 28.44 | Not recorded | 69a66c122d603e6de3e3cd5e | | AceMath-RM-7B | 26.72 | Not recorded | 69a66c122d603e6de3e3cd5e | | Math-Shepherd-7B | 26.04 | Not recorded | 69a66c122d603e6de3e3cd5e | ### OTV Weighted Majority Voting Accuracy: 69.32. Result ID: 69a82b8a6dbb7075b94e772d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### DeepConf Weighted Majority Voting Accuracy: 66.77. Result ID: 69a82b8a6dbb7075b94e7726. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### GenRM Weighted Majority Voting Accuracy: 66.72. Result ID: 69a82b8a6dbb7075b94e7727. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### VersaPRM-8B Weighted Majority Voting Accuracy: 66.72. Result ID: 69a82b8a6dbb7075b94e7729. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### Math-Shepherd-7B Weighted Majority Voting Accuracy: 66.56. Result ID: 69a82b8a6dbb7075b94e772a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### MAJ@128 Weighted Majority Voting Accuracy: 66.46. Result ID: 69a82b8a6dbb7075b94e7725. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### Qwen2.5-PRM800K-7B Weighted Majority Voting Accuracy: 66.35. Result ID: 69a82b8a6dbb7075b94e772b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### Qwen2.5-PRM-7B Weighted Majority Voting Accuracy: 63.49. Result ID: 69a82b8a6dbb7075b94e772c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### AceMath-RM-7B Weighted Majority Voting Accuracy: 60. Result ID: 69a82b8a6dbb7075b94e7728. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-4B"}. This record does not assert independent reproduction. ### OTV Weighted Majority Voting Accuracy: 49.58. Result ID: 69a82b8a6dbb7075b94e7743. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### Qwen2.5-PRM800K-7B Weighted Majority Voting Accuracy: 48.18. Result ID: 69a82b8a6dbb7075b94e7741. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### Qwen2.5-PRM-7B Weighted Majority Voting Accuracy: 46.77. Result ID: 69a82b8a6dbb7075b94e7742. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### DeepConf Weighted Majority Voting Accuracy: 44.79. Result ID: 69a82b8a6dbb7075b94e773c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### Math-Shepherd-7B Weighted Majority Voting Accuracy: 43.65. Result ID: 69a82b8a6dbb7075b94e7740. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### VersaPRM-8B Weighted Majority Voting Accuracy: 43.18. Result ID: 69a82b8a6dbb7075b94e773f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### AceMath-RM-7B Weighted Majority Voting Accuracy: 42.76. Result ID: 69a82b8a6dbb7075b94e773e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### GenRM Weighted Majority Voting Accuracy: 42.55. Result ID: 69a82b8a6dbb7075b94e773d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### MAJ@128 Weighted Majority Voting Accuracy: 41.77. Result ID: 69a82b8a6dbb7075b94e773b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"DAPO-QWEN-32B"}. This record does not assert independent reproduction. ### OTV Weighted Majority Voting Accuracy: 33.85. Result ID: 69a82b8a6dbb7075b94e7738. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### GenRM Weighted Majority Voting Accuracy: 32.66. Result ID: 69a82b8a6dbb7075b94e7732. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### DeepConf Weighted Majority Voting Accuracy: 31.87. Result ID: 69a82b8a6dbb7075b94e7731. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### Qwen2.5-PRM-7B Weighted Majority Voting Accuracy: 29.48. Result ID: 69a82b8a6dbb7075b94e7737. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### Qwen2.5-PRM800K-7B Weighted Majority Voting Accuracy: 29.32. Result ID: 69a82b8a6dbb7075b94e7736. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### VersaPRM-8B Weighted Majority Voting Accuracy: 29.11. Result ID: 69a82b8a6dbb7075b94e7734. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### MAJ@128 Weighted Majority Voting Accuracy: 28.44. Result ID: 69a82b8a6dbb7075b94e7730. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### AceMath-RM-7B Weighted Majority Voting Accuracy: 26.72. Result ID: 69a82b8a6dbb7075b94e7733. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. ### Math-Shepherd-7B Weighted Majority Voting Accuracy: 26.04. Result ID: 69a82b8a6dbb7075b94e7735. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"QWEN3-8B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.