# Multi-modal Visual Reasoning on VKnowU Source: [Multi-modal Visual Reasoning on VKnowU]() Collection index: [llms.txt]() Data updated: 2026-07-07T11:49:49.254Z Evaluation of model performance on the VKnowU dataset for Multi-modal Visual Reasoning, covering sub-tasks such as Intuitive Physics, Object Affordance, and Social Relations. 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 | Multi-modal Visual Reasoning | | Dataset (including any stated split) | VKnowU | | Metric | Overall Accuracy | | Unit | PERCENT | | Direction | Higher is better | | Results loaded | 30 of 30 | | 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 - [Overall Accuracy](): Higher is better; unit: PERCENT. - [Accuracy (IP)](): Higher is better; unit: PERCENT. - [Accuracy (OA)](): Higher is better; unit: PERCENT. - [Accuracy (OM)](): Higher is better; unit: PERCENT. - [Accuracy (SA)](): Higher is better; unit: PERCENT. - [Accuracy (WC)](): Higher is better; unit: PERCENT. - [Accuracy (EA)](): Higher is better; unit: PERCENT. - [Accuracy (MS)](): Higher is better; unit: PERCENT. - [Accuracy (SR)](): Higher is better; unit: PERCENT. - [Accuracy (SI)](): Higher is better; unit: PERCENT. - [Accuracy (HC)](): Higher is better; unit: PERCENT. ## Recorded results | Method | Overall Accuracy | Result date | Report paper | | --- | --- | --- | --- | | Human Performance | 93 | Not recorded | 692689a5d6b0dd33429304dd | | Gemini-3.0-Pro | 76.3 | Not recorded | 692689a5d6b0dd33429304dd | | Gemini-3.0-Flash | 76.1 | Not recorded | 692689a5d6b0dd33429304dd | | InternVL3.5-241B-A28B | 74.6 | Not recorded | 692689a5d6b0dd33429304dd | | InternVL3.5-38B | 72.7 | Not recorded | 692689a5d6b0dd33429304dd | | InternVL3.5-38B-Think | 71.9 | Not recorded | 692689a5d6b0dd33429304dd | | InternVL3.5-30B-A3B | 71.5 | Not recorded | 692689a5d6b0dd33429304dd | | Gemini-2.5-Pro | 71.1 | Not recorded | 692689a5d6b0dd33429304dd | | Qwen2.5-VL-72B-Instruct | 70.1 | Not recorded | 692689a5d6b0dd33429304dd | | LLaVA-Video-72B | 69.9 | Not recorded | 692689a5d6b0dd33429304dd | | Qwen3-VL-8B-Instruct | 69 | Not recorded | 692689a5d6b0dd33429304dd | | Gemini-2.5-Flash | 68.8 | Not recorded | 692689a5d6b0dd33429304dd | | GLM-4.6V-Flash | 68.4 | Not recorded | 692689a5d6b0dd33429304dd | | MiMo-VL-7B-RL | 68 | Not recorded | 692689a5d6b0dd33429304dd | | GLM-4.1V-9B-Thinking | 67.6 | Not recorded | 692689a5d6b0dd33429304dd | | Qwen3-VL-4B-Instruct | 67.6 | Not recorded | 692689a5d6b0dd33429304dd | | InternVL3.5-8B-Think | 67.6 | Not recorded | 692689a5d6b0dd33429304dd | | LLaVA-OV-72B | 66.9 | Not recorded | 692689a5d6b0dd33429304dd | | InternVL3.5-8B | 66.9 | Not recorded | 692689a5d6b0dd33429304dd | | GPT-5-mini | 66.7 | Not recorded | 692689a5d6b0dd33429304dd | | MiniCPM-V 4.5 | 65.9 | Not recorded | 692689a5d6b0dd33429304dd | | GPT-4o | 65.7 | Not recorded | 692689a5d6b0dd33429304dd | | LLaVA-Video-7B | 64.7 | Not recorded | 692689a5d6b0dd33429304dd | | mPLUG-Owl3-7B | 64.2 | Not recorded | 692689a5d6b0dd33429304dd | | Qwen2.5-VL-7B-Instruct | 64 | Not recorded | 692689a5d6b0dd33429304dd | | LLaVA-OV-7B | 63.3 | Not recorded | 692689a5d6b0dd33429304dd | | MiniCPM-V 2.6 | 62.3 | Not recorded | 692689a5d6b0dd33429304dd | | Qwen2.5-VL-3B-Instruct | 61.1 | Not recorded | 692689a5d6b0dd33429304dd | | VideoLLaMA2-7B | 55.7 | Not recorded | 692689a5d6b0dd33429304dd | | Random Guess | 37.4 | Not recorded | 692689a5d6b0dd33429304dd | ### Human Performance Overall Accuracy: 93. Result ID: 6a4ce7cd60b88bb339cbf54b. [Reporting paper]() Recorded evaluation configuration: {"Type":"Reference"}. This record does not assert independent reproduction. ### Gemini-3.0-Pro Overall Accuracy: 76.3. Result ID: 6a4ce7cd60b88bb339cbf567. [Reporting paper]() Recorded evaluation configuration: {"Access":"Proprietary"}. This record does not assert independent reproduction. ### Gemini-3.0-Flash Overall Accuracy: 76.1. Result ID: 6a4ce7cd60b88bb339cbf566. [Reporting paper]() Recorded evaluation configuration: {"Access":"Proprietary"}. This record does not assert independent reproduction. ### InternVL3.5-241B-A28B Overall Accuracy: 74.6. Result ID: 6a4ce7cd60b88bb339cbf561. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"241B"}. This record does not assert independent reproduction. ### InternVL3.5-38B Overall Accuracy: 72.7. Result ID: 6a4ce7cd60b88bb339cbf55f. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"38B"}. This record does not assert independent reproduction. ### InternVL3.5-38B-Think Overall Accuracy: 71.9. Result ID: 6a4ce7cd60b88bb339cbf560. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"38B"}. This record does not assert independent reproduction. ### InternVL3.5-30B-A3B Overall Accuracy: 71.5. Result ID: 6a4ce7cd60b88bb339cbf55e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"30B"}. This record does not assert independent reproduction. ### Gemini-2.5-Pro Overall Accuracy: 71.1. Result ID: 6a4ce7cd60b88bb339cbf565. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Proprietary"}. This record does not assert independent reproduction. ### Qwen2.5-VL-72B-Instruct Overall Accuracy: 70.1. Result ID: 6a4ce7cd60b88bb339cbf556. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"72B"}. This record does not assert independent reproduction. ### LLaVA-Video-72B Overall Accuracy: 69.9. Result ID: 6a4ce7cd60b88bb339cbf553. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"72B"}. This record does not assert independent reproduction. ### Qwen3-VL-8B-Instruct Overall Accuracy: 69. Result ID: 6a4ce7cd60b88bb339cbf55b. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"8B"}. This record does not assert independent reproduction. ### Gemini-2.5-Flash Overall Accuracy: 68.8. Result ID: 6a4ce7cd60b88bb339cbf564. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Proprietary"}. This record does not assert independent reproduction. ### GLM-4.6V-Flash Overall Accuracy: 68.4. Result ID: 6a4ce7cd60b88bb339cbf559. [Reporting paper]() Recorded evaluation configuration: {"Access":"Open-Source"}. This record does not assert independent reproduction. ### MiMo-VL-7B-RL Overall Accuracy: 68. Result ID: 6a4ce7cd60b88bb339cbf557. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"7B"}. This record does not assert independent reproduction. ### GLM-4.1V-9B-Thinking Overall Accuracy: 67.6. Result ID: 6a4ce7cd60b88bb339cbf558. [Reporting paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"9B"}. This record does not assert independent reproduction. ### Qwen3-VL-4B-Instruct Overall Accuracy: 67.6. Result ID: 6a4ce7cd60b88bb339cbf55a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"4B"}. This record does not assert independent reproduction. ### InternVL3.5-8B-Think Overall Accuracy: 67.6. Result ID: 6a4ce7cd60b88bb339cbf55d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"8B"}. This record does not assert independent reproduction. ### LLaVA-OV-72B Overall Accuracy: 66.9. Result ID: 6a4ce7cd60b88bb339cbf551. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"72B"}. This record does not assert independent reproduction. ### InternVL3.5-8B Overall Accuracy: 66.9. Result ID: 6a4ce7cd60b88bb339cbf55c. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"8B"}. This record does not assert independent reproduction. ### GPT-5-mini Overall Accuracy: 66.7. Result ID: 6a4ce7cd60b88bb339cbf563. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Proprietary"}. This record does not assert independent reproduction. ### MiniCPM-V 4.5 Overall Accuracy: 65.9. Result ID: 6a4ce7cd60b88bb339cbf54e. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source"}. This record does not assert independent reproduction. ### GPT-4o Overall Accuracy: 65.7. Result ID: 6a4ce7cd60b88bb339cbf562. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Proprietary"}. This record does not assert independent reproduction. ### LLaVA-Video-7B Overall Accuracy: 64.7. Result ID: 6a4ce7cd60b88bb339cbf552. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"7B"}. This record does not assert independent reproduction. ### mPLUG-Owl3-7B Overall Accuracy: 64.2. Result ID: 6a4ce7cd60b88bb339cbf54f. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"7B"}. This record does not assert independent reproduction. ### Qwen2.5-VL-7B-Instruct Overall Accuracy: 64. Result ID: 6a4ce7cd60b88bb339cbf555. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"7B"}. This record does not assert independent reproduction. ### LLaVA-OV-7B Overall Accuracy: 63.3. Result ID: 6a4ce7cd60b88bb339cbf550. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"7B"}. This record does not assert independent reproduction. ### MiniCPM-V 2.6 Overall Accuracy: 62.3. Result ID: 6a4ce7cd60b88bb339cbf54d. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source"}. This record does not assert independent reproduction. ### Qwen2.5-VL-3B-Instruct Overall Accuracy: 61.1. Result ID: 6a4ce7cd60b88bb339cbf554. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"3B"}. This record does not assert independent reproduction. ### VideoLLaMA2-7B Overall Accuracy: 55.7. Result ID: 6a4ce7cd60b88bb339cbf54c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Access":"Open-Source","Size":"7B"}. This record does not assert independent reproduction. ### Random Guess Overall Accuracy: 37.4. Result ID: 6a4ce7cd60b88bb339cbf54a. [Reporting paper]() Recorded evaluation configuration: {"Type":"Reference"}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.