# Multimodal Mathematical Reasoning on WeMath (test) Source: [Multimodal Mathematical Reasoning on WeMath (test)]() Collection index: [llms.txt]() Data updated: 2026-03-31T08:40:59.324Z Evaluation of model performance on the test split of the WeMath dataset for Multimodal Mathematical Reasoning, typically measured by Accuracy and Average score. 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 | Multimodal Mathematical Reasoning | | Dataset (including any stated split) | WeMath (test) | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 17 of 17 | | 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. - [Average Score](): Higher is better; unit: SCALAR. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | GSPO + KAWHI (Qwen3) | 72.15 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GSPO (Qwen3) | 71.55 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | DAPO + KAWHI (Qwen3) | 71.23 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GRPO + KAWHI (Qwen3) | 70.98 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GRPO (Qwen3) | 70.31 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | DAPO (Qwen3) | 69.67 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | Qwen3-VL-4B-Instruct Base | 69.43 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GSPO + KAWHI | 59.71 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GSPO | 59.02 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | DAPO + KAWHI | 58.21 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GRPO | 58.1 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | DAPO | 57.98 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | GRPO + KAWHI | 57.84 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | VPPO | 57.12 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | FT-RL | 56.78 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | Step-GRPO | 55.32 | Not recorded | 69cb54fe0636e5bf4afc9e86 | | Qwen2.5-VL-7B-Instruct Base | 54.32 | Not recorded | 69cb54fe0636e5bf4afc9e86 | ### GSPO + KAWHI (Qwen3) Accuracy: 72.15. Result ID: 69cb888a47f34ce924b3dfe9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"uniform reward","RL Method":"KAWHI"}. This record does not assert independent reproduction. ### GSPO (Qwen3) Accuracy: 71.55. Result ID: 69cb888a47f34ce924b3dfe8. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"uniform reward","RL Method":"GSPO"}. This record does not assert independent reproduction. ### DAPO + KAWHI (Qwen3) Accuracy: 71.23. Result ID: 69cb888a47f34ce924b3dfe7. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"uniform reward","RL Method":"KAWHI"}. This record does not assert independent reproduction. ### GRPO + KAWHI (Qwen3) Accuracy: 70.98. Result ID: 69cb888a47f34ce924b3dfe5. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"uniform reward","RL Method":"KAWHI"}. This record does not assert independent reproduction. ### GRPO (Qwen3) Accuracy: 70.31. Result ID: 69cb888a47f34ce924b3dfe4. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"uniform reward","RL Method":"GRPO"}. This record does not assert independent reproduction. ### DAPO (Qwen3) Accuracy: 69.67. Result ID: 69cb888a47f34ce924b3dfe6. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"uniform reward","RL Method":"DAPO"}. This record does not assert independent reproduction. ### Qwen3-VL-4B-Instruct Base Accuracy: 69.43. Result ID: 69cb888a47f34ce924b3dfe3. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen3-VL-4B-Instruct","Reward Strategy":"none","RL Method":"Base"}. This record does not assert independent reproduction. ### GSPO + KAWHI Accuracy: 59.71. Result ID: 69cb888a47f34ce924b3dfe2. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"uniform reward","RL Method":"KAWHI"}. This record does not assert independent reproduction. ### GSPO Accuracy: 59.02. Result ID: 69cb888a47f34ce924b3dfe1. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"uniform reward","RL Method":"GSPO"}. This record does not assert independent reproduction. ### DAPO + KAWHI Accuracy: 58.21. Result ID: 69cb888a47f34ce924b3dfe0. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"uniform reward","RL Method":"KAWHI"}. This record does not assert independent reproduction. ### GRPO Accuracy: 58.1. Result ID: 69cb888a47f34ce924b3dfdd. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"uniform reward","RL Method":"GRPO"}. This record does not assert independent reproduction. ### DAPO Accuracy: 57.98. Result ID: 69cb888a47f34ce924b3dfdf. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"uniform reward","RL Method":"DAPO"}. This record does not assert independent reproduction. ### GRPO + KAWHI Accuracy: 57.84. Result ID: 69cb888a47f34ce924b3dfde. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"uniform reward","RL Method":"KAWHI"}. This record does not assert independent reproduction. ### VPPO Accuracy: 57.12. Result ID: 69cb888a47f34ce924b3dfdc. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"fine-grained reward","RL Method":"VPPO"}. This record does not assert independent reproduction. ### FT-RL Accuracy: 56.78. Result ID: 69cb888a47f34ce924b3dfdb. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"fine-grained reward","RL Method":"FT-RL"}. This record does not assert independent reproduction. ### Step-GRPO Accuracy: 55.32. Result ID: 69cb888a47f34ce924b3dfda. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"fine-grained reward","RL Method":"Step-GRPO"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B-Instruct Base Accuracy: 54.32. Result ID: 69cb888a47f34ce924b3dfd9. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone Model":"Qwen2.5-VL-7B-Instruct","Reward Strategy":"none","RL Method":"Base"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.