# Mathematical Reasoning on AIME 2024 (AVG@32, Pass@32) Source: [Mathematical Reasoning on AIME 2024 (AVG@32, Pass@32)]() Collection index: [llms.txt]() Data updated: 2026-07-07T08:12:03.689Z Evaluation of model performance on the American Invitational Mathematics Examination (AIME) 2024 dataset, reporting results using AVG@32 and Pass@32 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 2024 | | Metric | AVG@32 | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 12 of 12 | | 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 - [AVG@32](): Higher is better; unit: PERCENT. - [Pass@32](): Higher is better; unit: PERCENT. ## Recorded results | Method | AVG@32 | Result date | Report paper | | --- | --- | --- | --- | | OLMo-3.1-32B-Think | 79.56 | Not recorded | 6a4c87e085dba69fef8b238b | | Projected (Ours) | 78.88 | Not recorded | 6a4c87e085dba69fef8b238b | | Polaris | 78.75 | Not recorded | 6a4c87e085dba69fef8b238b | | Projected (Ours) | 78.02 | Not recorded | 6a4c87e085dba69fef8b238b | | OLMo3-32B-think-dpo | 74.48 | Not recorded | 6a4c87e085dba69fef8b238b | | Qwen3 4B | 72.5 | Not recorded | 6a4c87e085dba69fef8b238b | | OLMo-3.1-7B-RL-Zero-Math | 50.21 | Not recorded | 6a4c87e085dba69fef8b238b | | Projected (Ours) | 49.02 | Not recorded | 6a4c87e085dba69fef8b238b | | DeepscaleR | 40.31 | Not recorded | 6a4c87e085dba69fef8b238b | | Projected (Ours) | 40.21 | Not recorded | 6a4c87e085dba69fef8b238b | | Deepseek-distill-qwen2.5 | 31.67 | Not recorded | 6a4c87e085dba69fef8b238b | | OLMo3-7B-Base | 22.08 | Not recorded | 6a4c87e085dba69fef8b238b | ### OLMo-3.1-32B-Think AVG@32: 79.56. Result ID: 6a4cb4b860b88bb339c83f96. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"32B","Training Recipe Category":"The SOTA Training Recipes","Type":"RL"}. This record does not assert independent reproduction. ### Projected (Ours) AVG@32: 78.88. Result ID: 6a4cb4b860b88bb339c83f97. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"32B","Training Recipe Category":"The SOTA Training Recipes","Rank":"1%"}. This record does not assert independent reproduction. ### Polaris AVG@32: 78.75. Result ID: 6a4cb4b860b88bb339c83f93. [Reporting paper]() Recorded evaluation configuration: {"Model Size":"4B","Training Recipe Category":"The SOTA Training Recipes","Type":"RL"}. This record does not assert independent reproduction. ### Projected (Ours) AVG@32: 78.02. Result ID: 6a4cb4b860b88bb339c83f94. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"4B","Training Recipe Category":"The SOTA Training Recipes","Rank":"10%"}. This record does not assert independent reproduction. ### OLMo3-32B-think-dpo AVG@32: 74.48. Result ID: 6a4cb4b860b88bb339c83f95. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"32B","Training Recipe Category":"The SOTA Training Recipes","Type":"Base"}. This record does not assert independent reproduction. ### Qwen3 4B AVG@32: 72.5. Result ID: 6a4cb4b860b88bb339c83f92. [Reporting paper]() Recorded evaluation configuration: {"Model Size":"4B","Training Recipe Category":"The SOTA Training Recipes","Type":"Base"}. This record does not assert independent reproduction. ### OLMo-3.1-7B-RL-Zero-Math AVG@32: 50.21. Result ID: 6a4cb4b860b88bb339c83f99. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"7B","Training Recipe Category":"Emergent Reasoning Ability from the Base Model","Type":"RL"}. This record does not assert independent reproduction. ### Projected (Ours) AVG@32: 49.02. Result ID: 6a4cb4b860b88bb339c83f9a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"7B","Training Recipe Category":"Emergent Reasoning Ability from the Base Model","Rank":"30%"}. This record does not assert independent reproduction. ### DeepscaleR AVG@32: 40.31. Result ID: 6a4cb4b860b88bb339c83f90. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"1.5B","Training Recipe Category":"The SOTA Training Recipes","Type":"RL"}. This record does not assert independent reproduction. ### Projected (Ours) AVG@32: 40.21. Result ID: 6a4cb4b860b88bb339c83f91. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"1.5B","Training Recipe Category":"The SOTA Training Recipes","Rank":"1%"}. This record does not assert independent reproduction. ### Deepseek-distill-qwen2.5 AVG@32: 31.67. Result ID: 6a4cb4b860b88bb339c83f8f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"1.5B","Training Recipe Category":"The SOTA Training Recipes","Type":"Base"}. This record does not assert independent reproduction. ### OLMo3-7B-Base AVG@32: 22.08. Result ID: 6a4cb4b860b88bb339c83f98. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Size":"7B","Training Recipe Category":"Emergent Reasoning Ability from the Base Model","Type":"Base"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.