# Mathematical Reasoning on MATH (MH) Source: [Mathematical Reasoning on MATH (MH)]() Collection index: [llms.txt]() Data updated: 2026-05-13T09:54:18.862Z Evaluation of model performance on the MATH dataset for mathematical reasoning, reporting the Mean Harmonic Score (MH). 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 | | Metric | Mean Harmonic Score (MH) | | 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 - [Mean Harmonic Score (MH)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Mean Harmonic Score (MH) | Result date | Report paper | | --- | --- | --- | --- | | GEAR | 80.2 | Not recorded | 6a04059cdbc836b87d82758f | | OPSD | 78.3 | Not recorded | 6a04059cdbc836b87d82758f | | GEAR | 77.6 | Not recorded | 6a04059cdbc836b87d82758f | | ARPO | 76.4 | Not recorded | 6a04059cdbc836b87d82758f | | MT-GRPO | 76.2 | Not recorded | 6a04059cdbc836b87d82758f | | GRPO | 75.8 | Not recorded | 6a04059cdbc836b87d82758f | | GRPO | 75.4 | Not recorded | 6a04059cdbc836b87d82758f | | OPSD+RL | 74.6 | Not recorded | 6a04059cdbc836b87d82758f | | ARPO | 74.2 | Not recorded | 6a04059cdbc836b87d82758f | | GEAR | 74 | Not recorded | 6a04059cdbc836b87d82758f | | OPSD | 72.8 | Not recorded | 6a04059cdbc836b87d82758f | | GRPO | 72.8 | Not recorded | 6a04059cdbc836b87d82758f | | Base | 72.5 | Not recorded | 6a04059cdbc836b87d82758f | | MT-GRPO | 72.4 | Not recorded | 6a04059cdbc836b87d82758f | | Base | 72.1 | Not recorded | 6a04059cdbc836b87d82758f | | OPSD+RL | 71.9 | Not recorded | 6a04059cdbc836b87d82758f | ### GEAR Mean Harmonic Score (MH): 80.2. Result ID: 6a0448eb47f34ce924ea9340. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### OPSD Mean Harmonic Score (MH): 78.3. Result ID: 6a0448eb47f34ce924ea933b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### GEAR Mean Harmonic Score (MH): 77.6. Result ID: 6a0448eb47f34ce924ea9349. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Multi-Domain Generalization"}. This record does not assert independent reproduction. ### ARPO Mean Harmonic Score (MH): 76.4. Result ID: 6a0448eb47f34ce924ea933e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### MT-GRPO Mean Harmonic Score (MH): 76.2. Result ID: 6a0448eb47f34ce924ea933f. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### GRPO Mean Harmonic Score (MH): 75.8. Result ID: 6a0448eb47f34ce924ea9348. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Multi-Domain Generalization"}. This record does not assert independent reproduction. ### GRPO Mean Harmonic Score (MH): 75.4. Result ID: 6a0448eb47f34ce924ea933d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### OPSD+RL Mean Harmonic Score (MH): 74.6. Result ID: 6a0448eb47f34ce924ea933c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### ARPO Mean Harmonic Score (MH): 74.2. Result ID: 6a0448eb47f34ce924ea9345. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### GEAR Mean Harmonic Score (MH): 74. Result ID: 6a0448eb47f34ce924ea9347. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### OPSD Mean Harmonic Score (MH): 72.8. Result ID: 6a0448eb47f34ce924ea9342. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### GRPO Mean Harmonic Score (MH): 72.8. Result ID: 6a0448eb47f34ce924ea9344. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### Base Mean Harmonic Score (MH): 72.5. Result ID: 6a0448eb47f34ce924ea933a. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-4B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### MT-GRPO Mean Harmonic Score (MH): 72.4. Result ID: 6a0448eb47f34ce924ea9346. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### Base Mean Harmonic Score (MH): 72.1. Result ID: 6a0448eb47f34ce924ea9341. [Reporting paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. ### OPSD+RL Mean Harmonic Score (MH): 71.9. Result ID: 6a0448eb47f34ce924ea9343. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen3-8B","Evaluation Context":"Standard"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.