# Mathematics on AMC 23 (avg@32 accuracy) Source: [Mathematics on AMC 23 (avg@32 accuracy)]() Collection index: [llms.txt]() Data updated: 2026-05-28T08:54:10.751Z Evaluation of model performance on the AMC 23 dataset for general Mathematics tasks, measured using the avg@32 accuracy metric. 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 | Mathematics | | Dataset (including any stated split) | AMC 23 | | Metric | Avg@32 Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 16 of 16 | | 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 Accuracy](): Higher is better; unit: PERCENT. ## Recorded results | Method | Avg@32 Accuracy | Result date | Report paper | | --- | --- | --- | --- | | IBTPO | 57.9 | Not recorded | 6a17cc00dbd7150506f81233 | | TreeRL | 56.1 | Not recorded | 6a17cc00dbd7150506f81233 | | TreePO | 55.8 | Not recorded | 6a17cc00dbd7150506f81233 | | IBRO | 55.3 | Not recorded | 6a17cc00dbd7150506f81233 | | GRPO w/ Entropy Reg† | 54.3 | Not recorded | 6a17cc00dbd7150506f81233 | | Vanilla GRPO | 53.5 | Not recorded | 6a17cc00dbd7150506f81233 | | GRPO w/ Clip-higher | 53.3 | Not recorded | 6a17cc00dbd7150506f81233 | | Initial Model | 41.7 | Not recorded | 6a17cc00dbd7150506f81233 | | IBTPO | 38.7 | Not recorded | 6a17cc00dbd7150506f81233 | | TreeRL | 37 | Not recorded | 6a17cc00dbd7150506f81233 | | TreePO | 36.5 | Not recorded | 6a17cc00dbd7150506f81233 | | Vanilla GRPO | 35.6 | Not recorded | 6a17cc00dbd7150506f81233 | | GRPO w/ Clip-higher | 35 | Not recorded | 6a17cc00dbd7150506f81233 | | IBRO | 34.6 | Not recorded | 6a17cc00dbd7150506f81233 | | GRPO w/ Entropy Reg† | 33.6 | Not recorded | 6a17cc00dbd7150506f81233 | | Initial Model | 24.5 | Not recorded | 6a17cc00dbd7150506f81233 | ### IBTPO Avg@32 Accuracy: 57.9. Result ID: 6a180298fbc373592c11dc11. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### TreeRL Avg@32 Accuracy: 56.1. Result ID: 6a180298fbc373592c11dc0f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### TreePO Avg@32 Accuracy: 55.8. Result ID: 6a180298fbc373592c11dc10. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### IBRO Avg@32 Accuracy: 55.3. Result ID: 6a180298fbc373592c11dc0e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### GRPO w/ Entropy Reg† Avg@32 Accuracy: 54.3. Result ID: 6a180298fbc373592c11dc0d. [Reporting paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### Vanilla GRPO Avg@32 Accuracy: 53.5. Result ID: 6a180298fbc373592c11dc0b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### GRPO w/ Clip-higher Avg@32 Accuracy: 53.3. Result ID: 6a180298fbc373592c11dc0c. [Reporting paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### Initial Model Avg@32 Accuracy: 41.7. Result ID: 6a180298fbc373592c11dc0a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-8B-Base"}. This record does not assert independent reproduction. ### IBTPO Avg@32 Accuracy: 38.7. Result ID: 6a180298fbc373592c11dc09. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### TreeRL Avg@32 Accuracy: 37. Result ID: 6a180298fbc373592c11dc07. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### TreePO Avg@32 Accuracy: 36.5. Result ID: 6a180298fbc373592c11dc08. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### Vanilla GRPO Avg@32 Accuracy: 35.6. Result ID: 6a180298fbc373592c11dc03. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### GRPO w/ Clip-higher Avg@32 Accuracy: 35. Result ID: 6a180298fbc373592c11dc04. [Reporting paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### IBRO Avg@32 Accuracy: 34.6. Result ID: 6a180298fbc373592c11dc06. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### GRPO w/ Entropy Reg† Avg@32 Accuracy: 33.6. Result ID: 6a180298fbc373592c11dc05. [Reporting paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. ### Initial Model Avg@32 Accuracy: 24.5. Result ID: 6a180298fbc373592c11dc02. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Base Model":"Qwen3-1.7B-Base"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.