# Mathematical Reasoning on MathVista MVistam Source: [Mathematical Reasoning on MathVista MVistam]() Collection index: [llms.txt]() Data updated: 2026-05-05T06:54:29.505Z Evaluation of model performance on the MathVista MVistam dataset, focusing on general mathematical reasoning capabilities, typically measured by Accuracy. 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) | MathVista MVistam | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 36 of 36 | | 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. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Gemini-2.0-Flash | 73.4 | Not recorded | 695b456053fdd5c361742edb | | CPPO | 72.2 | Not recorded | 695b456053fdd5c361742edb | | PAPO | 71.6 | Not recorded | 695b456053fdd5c361742edb | | GRPO | 71.2 | Not recorded | 695b456053fdd5c361742edb | | NoisyRollout | 71.1 | Not recorded | 695b456053fdd5c361742edb | | OpenVLThinker | 70.7 | Not recorded | 695b456053fdd5c361742edb | | PerceptionR1 | 70 | Not recorded | 695b456053fdd5c361742edb | | Look-Back | 69.1 | Not recorded | 695b456053fdd5c361742edb | | Vision-Matters | 68.6 | Not recorded | 695b456053fdd5c361742edb | | Vision-SR1 | 67 | Not recorded | 695b456053fdd5c361742edb | | CPPO | 66.3 | Not recorded | 695b456053fdd5c361742edb | | Qwen2.5-VL-7B | 65.6 | Not recorded | 695b456053fdd5c361742edb | | PAPO | 64.8 | Not recorded | 695b456053fdd5c361742edb | | CoVT | 64.3 | Not recorded | 69f979fbc0f56ed843089c9a | | GRPO | 63.7 | Not recorded | 695b456053fdd5c361742edb | | Unsilencing Latent Reasoning | 63.7 | Not recorded | 69f979fbc0f56ed843089c9a | | LVRRF | 62.6 | Not recorded | 69f979fbc0f56ed843089c9a | | Monet | 62.4 | Not recorded | 69f979fbc0f56ed843089c9a | | Visionary-R1 | 61.4 | Not recorded | 695b456053fdd5c361742edb | | LVR | 61.4 | Not recorded | 69f979fbc0f56ed843089c9a | | GPT4-o | 60 | Not recorded | 695b456053fdd5c361742edb | | OpenVLThinker | 60 | Not recorded | 695b456053fdd5c361742edb | | DMLR | 59.1 | Not recorded | 69f979fbc0f56ed843089c9a | | ICoT | 58.9 | Not recorded | 69f979fbc0f56ed843089c9a | | Vanilla | 58.7 | Not recorded | 69f979fbc0f56ed843089c9a | | CCoT | 57.8 | Not recorded | 69f979fbc0f56ed843089c9a | | Qwen2.5-VL-3B | 56.4 | Not recorded | 695b456053fdd5c361742edb | | MCoT | 56.4 | Not recorded | 69f979fbc0f56ed843089c9a | | DMLR | 51 | Not recorded | 69f979fbc0f56ed843089c9a | | Unsilencing Latent Reasoning | 50 | Not recorded | 69f979fbc0f56ed843089c9a | | ICoT | 49.8 | Not recorded | 69f979fbc0f56ed843089c9a | | LVRRF | 48.6 | Not recorded | 69f979fbc0f56ed843089c9a | | Vanilla | 48.2 | Not recorded | 69f979fbc0f56ed843089c9a | | CCoT | 48 | Not recorded | 69f979fbc0f56ed843089c9a | | LVR | 47.5 | Not recorded | 69f979fbc0f56ed843089c9a | | MCoT | 47.3 | Not recorded | 69f979fbc0f56ed843089c9a | ### Gemini-2.0-Flash Accuracy: 73.4. Result ID: 6997f3ea5698665ec44d5564. [Reporting paper]() Recorded evaluation configuration: {"Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### CPPO Accuracy: 72.2. Result ID: 6997f3ea5698665ec44d5574. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### PAPO Accuracy: 71.6. Result ID: 6997f3ea5698665ec44d5570. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### GRPO Accuracy: 71.2. Result ID: 6997f3ea5698665ec44d5573. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### NoisyRollout Accuracy: 71.1. Result ID: 6997f3ea5698665ec44d5572. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### OpenVLThinker Accuracy: 70.7. Result ID: 6997f3ea5698665ec44d556c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### PerceptionR1 Accuracy: 70. Result ID: 6997f3ea5698665ec44d5571. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### Look-Back Accuracy: 69.1. Result ID: 6997f3ea5698665ec44d556e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### Vision-Matters Accuracy: 68.6. Result ID: 6997f3ea5698665ec44d556f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### Vision-SR1 Accuracy: 67. Result ID: 6997f3ea5698665ec44d556d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### CPPO Accuracy: 66.3. Result ID: 6997f3ea5698665ec44d556a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B Accuracy: 65.6. Result ID: 6997f3ea5698665ec44d556b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-7B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### PAPO Accuracy: 64.8. Result ID: 6997f3ea5698665ec44d5568. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### CoVT Accuracy: 64.3. Result ID: 69f9941e47f34ce924dcc210. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### GRPO Accuracy: 63.7. Result ID: 6997f3ea5698665ec44d5569. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### Unsilencing Latent Reasoning Accuracy: 63.7. Result ID: 69f9941e47f34ce924dcc213. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### LVRRF Accuracy: 62.6. Result ID: 69f9941e47f34ce924dcc212. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### Monet Accuracy: 62.4. Result ID: 69f9941e47f34ce924dcc20f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### Visionary-R1 Accuracy: 61.4. Result ID: 6997f3ea5698665ec44d5567. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### LVR Accuracy: 61.4. Result ID: 69f9941e47f34ce924dcc211. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### GPT4-o Accuracy: 60. Result ID: 6997f3ea5698665ec44d5563. [Reporting paper]() Recorded evaluation configuration: {"Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### OpenVLThinker Accuracy: 60. Result ID: 6997f3ea5698665ec44d5566. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### DMLR Accuracy: 59.1. Result ID: 69f9941e47f34ce924dcc20e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### ICoT Accuracy: 58.9. Result ID: 69f9941e47f34ce924dcc20d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 58.7. Result ID: 69f9941e47f34ce924dcc20a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### CCoT Accuracy: 57.8. Result ID: 69f9941e47f34ce924dcc20c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-3B Accuracy: 56.4. Result ID: 6997f3ea5698665ec44d5565. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5-VL-3B","Sampling Strategy":"avg@8","Inference Temperature":1}. This record does not assert independent reproduction. ### MCoT Accuracy: 56.4. Result ID: 69f9941e47f34ce924dcc20b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-7B"}. This record does not assert independent reproduction. ### DMLR Accuracy: 51. Result ID: 69f9941e47f34ce924dcc218. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### Unsilencing Latent Reasoning Accuracy: 50. Result ID: 69f9941e47f34ce924dcc21b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### ICoT Accuracy: 49.8. Result ID: 69f9941e47f34ce924dcc217. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### LVRRF Accuracy: 48.6. Result ID: 69f9941e47f34ce924dcc21a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### Vanilla Accuracy: 48.2. Result ID: 69f9941e47f34ce924dcc214. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### CCoT Accuracy: 48. Result ID: 69f9941e47f34ce924dcc216. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### LVR Accuracy: 47.5. Result ID: 69f9941e47f34ce924dcc219. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. ### MCoT Accuracy: 47.3. Result ID: 69f9941e47f34ce924dcc215. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Backbone":"Qwen2.5VL-3B"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.