# Medical Visual Question Answering on SLAKE (Acc, Recall, Rouge) Source: [Medical Visual Question Answering on SLAKE (Acc, Recall, Rouge)]() Collection index: [llms.txt]() Data updated: 2026-05-22T08:25:29.902Z Evaluation of model performance on the SLAKE dataset for Medical Visual Question Answering, reporting Accuracy, Recall, and Rouge scores. 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 | Medical Visual Question Answering | | Dataset (including any stated split) | SLAKE | | Metric | Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 25 of 25 | | 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. - [Recall](): Higher is better; unit: PERCENT. - [ROUGE Score](): Higher is better; unit: PERCENT. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Ours (Med-Evo) | 71.55 | Not recorded | 69afa5a62d603e6de3e43cef | | TTRL | 71.27 | Not recorded | 69afa5a62d603e6de3e43cef | | EN-INF | 68.73 | Not recorded | 69afa5a62d603e6de3e43cef | | TTRV | 68.45 | Not recorded | 69afa5a62d603e6de3e43cef | | DeepEyes-v2 | 66.2 | Not recorded | 6a0fe330f984739ed6739f4d | | MAESTRO | 66.2 | Not recorded | 6a0fe330f984739ed6739f4d | | GLM-4.6V | 63.1 | Not recorded | 6a0fe330f984739ed6739f4d | | Thyme | 62.6 | Not recorded | 6a0fe330f984739ed6739f4d | | GPT-5 | 61.8 | Not recorded | 6a0fe330f984739ed6739f4d | | Base Model | 61.41 | Not recorded | 69afa5a62d603e6de3e43cef | | VTOOL-R1 | 60.7 | Not recorded | 6a0fe330f984739ed6739f4d | | PixelReasoner | 59.8 | Not recorded | 6a0fe330f984739ed6739f4d | | Kimi-K2.5 | 59.6 | Not recorded | 6a0fe330f984739ed6739f4d | | DeepEyes | 58.7 | Not recorded | 6a0fe330f984739ed6739f4d | | GPT-4o | 58.5 | Not recorded | 6a0fe330f984739ed6739f4d | | Visual-ARFT | 58.5 | Not recorded | 6a0fe330f984739ed6739f4d | | VTS-V | 57.9 | Not recorded | 6a0fe330f984739ed6739f4d | | Qwen3-VL-32B | 57.6 | Not recorded | 6a0fe330f984739ed6739f4d | | Gemini-2.5-Pro | 56.8 | Not recorded | 6a0fe330f984739ed6739f4d | | VisionReasoner | 56.8 | Not recorded | 6a0fe330f984739ed6739f4d | | Gemini-2.5-Flash | 56 | Not recorded | 6a0fe330f984739ed6739f4d | | Direct Answering | 56 | Not recorded | 6a0fe330f984739ed6739f4d | | Untrained Model | 54.3 | Not recorded | 6a0fe330f984739ed6739f4d | | MathCoder-VL | 54 | Not recorded | 6a0fe330f984739ed6739f4d | | Chain-of-Focus | 48.2 | Not recorded | 6a0fe330f984739ed6739f4d | ### Ours (Med-Evo) Accuracy: 71.55. Result ID: 69afefcc6dbb7075b958ebad. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"base model":"MedVLM-R1"}. This record does not assert independent reproduction. ### TTRL Accuracy: 71.27. Result ID: 69afefcc6dbb7075b958ebac. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"base model":"MedVLM-R1"}. This record does not assert independent reproduction. ### EN-INF Accuracy: 68.73. Result ID: 69afefcc6dbb7075b958ebaa. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"base model":"MedVLM-R1"}. This record does not assert independent reproduction. ### TTRV Accuracy: 68.45. Result ID: 69afefcc6dbb7075b958ebab. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"base model":"MedVLM-R1"}. This record does not assert independent reproduction. ### DeepEyes-v2 Accuracy: 66.2. Result ID: 6a1012f4fbc373592c090597. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### MAESTRO Accuracy: 66.2. Result ID: 6a1012f4fbc373592c0905a0. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Type":"Ours"}. This record does not assert independent reproduction. ### GLM-4.6V Accuracy: 63.1. Result ID: 6a1012f4fbc373592c090591. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Open-Source & Baselines"}. This record does not assert independent reproduction. ### Thyme Accuracy: 62.6. Result ID: 6a1012f4fbc373592c090598. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### GPT-5 Accuracy: 61.8. Result ID: 6a1012f4fbc373592c09058e. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Closed-Source"}. This record does not assert independent reproduction. ### Base Model Accuracy: 61.41. Result ID: 69afefcc6dbb7075b958eba9. [Reporting paper]() Recorded evaluation configuration: {"base model":"MedVLM-R1"}. This record does not assert independent reproduction. ### VTOOL-R1 Accuracy: 60.7. Result ID: 6a1012f4fbc373592c090599. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### PixelReasoner Accuracy: 59.8. Result ID: 6a1012f4fbc373592c09059e. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### Kimi-K2.5 Accuracy: 59.6. Result ID: 6a1012f4fbc373592c090592. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Open-Source & Baselines"}. This record does not assert independent reproduction. ### DeepEyes Accuracy: 58.7. Result ID: 6a1012f4fbc373592c090596. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### GPT-4o Accuracy: 58.5. Result ID: 6a1012f4fbc373592c09058d. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Closed-Source"}. This record does not assert independent reproduction. ### Visual-ARFT Accuracy: 58.5. Result ID: 6a1012f4fbc373592c09059c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### VTS-V Accuracy: 57.9. Result ID: 6a1012f4fbc373592c09059a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### Qwen3-VL-32B Accuracy: 57.6. Result ID: 6a1012f4fbc373592c090593. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Open-Source & Baselines"}. This record does not assert independent reproduction. ### Gemini-2.5-Pro Accuracy: 56.8. Result ID: 6a1012f4fbc373592c090590. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Closed-Source"}. This record does not assert independent reproduction. ### VisionReasoner Accuracy: 56.8. Result ID: 6a1012f4fbc373592c09059d. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### Gemini-2.5-Flash Accuracy: 56. Result ID: 6a1012f4fbc373592c09058f. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Closed-Source"}. This record does not assert independent reproduction. ### Direct Answering Accuracy: 56. Result ID: 6a1012f4fbc373592c090594. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Open-Source & Baselines"}. This record does not assert independent reproduction. ### Untrained Model Accuracy: 54.3. Result ID: 6a1012f4fbc373592c090595. [Reporting paper]() Recorded evaluation configuration: {"Model Type":"Open-Source & Baselines"}. This record does not assert independent reproduction. ### MathCoder-VL Accuracy: 54. Result ID: 6a1012f4fbc373592c09059b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. ### Chain-of-Focus Accuracy: 48.2. Result ID: 6a1012f4fbc373592c09059f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Model Type":"Think with Images Methods"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.