# Vision-Language Evaluation on CV-Bench Source: [Vision-Language Evaluation on CV-Bench]() Collection index: [llms.txt]() Data updated: 2026-02-26T05:30:10.713Z Evaluation of models on the CV-Bench dataset focusing on general vision-language 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 | Vision-Language Evaluation | | Dataset (including any stated split) | CV-Bench | | 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. ## Recorded results | Method | Accuracy | Result date | Report paper | | --- | --- | --- | --- | | VLSI-2B | 90.1 | Not recorded | 68fa7c6e1e209131cea8e0ab | | VLSI-7B | 89.1 | Not recorded | 68fa7c6e1e209131cea8e0ab | | Cambrian-34B | 76.9 | Not recorded | 68fa7c6e1e209131cea8e0ab | | Phantom-7B | 74.9 | Not recorded | 68fa7c6e1e209131cea8e0ab | | Phantom | 74.9 | Not recorded | 68fa74ea1e209131cea64b95 | | Qwen2-VL-72B | 74.3 | Not recorded | 68fa7c6e1e209131cea8e0ab | | Florence-VL 8B | 73.4 | Not recorded | 68fa760f1e209131cea6b32d | | Cambrian-1 | 72.2 | Not recorded | 68fa74ea1e209131cea64b95 | | Cambrian 8B | 72.2 | Not recorded | 68fa760f1e209131cea6b32d | | Florence-VL 3B | 70.2 | Not recorded | 68fa760f1e209131cea6b32d | | Phi 3.5 Vision | 69.3 | Not recorded | 68fa760f1e209131cea6b32d | | GPT-4V (0409) | 69.1 | Not recorded | 68fa7c6e1e209131cea8e0ab | | GPT-4V | 69.1 | Not recorded | 68fa74ea1e209131cea64b95 | | LLaVA next 8B | 63.8 | Not recorded | 68fa760f1e209131cea6b32d | | Phantom-1.8B | 63.1 | Not recorded | 68fa7c6e1e209131cea8e0ab | | Mini-Gemini-HD 8B | 62.6 | Not recorded | 68fa760f1e209131cea6b32d | | Vila 3B | 55.2 | Not recorded | 68fa760f1e209131cea6b32d | ### VLSI-2B Accuracy: 90.1. Result ID: 698b91007a967eddb76fa30d. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"2B"}. This record does not assert independent reproduction. ### VLSI-7B Accuracy: 89.1. Result ID: 698b91007a967eddb76fa30e. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction. ### Cambrian-34B Accuracy: 76.9. Result ID: 698b91007a967eddb76fa308. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Size":"34B"}. This record does not assert independent reproduction. ### Phantom-7B Accuracy: 74.9. Result ID: 698b91007a967eddb76fa30c. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"7B"}. This record does not assert independent reproduction. ### Phantom Accuracy: 74.9. Result ID: 69936a983a6c31484915c852. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Parameters":"7B"}. This record does not assert independent reproduction. ### Qwen2-VL-72B Accuracy: 74.3. Result ID: 698b91007a967eddb76fa30a. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"72B"}. This record does not assert independent reproduction. ### Florence-VL 8B Accuracy: 73.4. Result ID: 69940cfd3a6c3148491ab68b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"# Vis tok.":576}. This record does not assert independent reproduction. ### Cambrian-1 Accuracy: 72.2. Result ID: 69936a983a6c31484915c850. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"Parameters":"8B"}. This record does not assert independent reproduction. ### Cambrian 8B Accuracy: 72.2. Result ID: 69940cfd3a6c3148491ab68a. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"# Vis tok.":576}. This record does not assert independent reproduction. ### Florence-VL 3B Accuracy: 70.2. Result ID: 69940cfd3a6c3148491ab687. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"# Vis tok.":576}. This record does not assert independent reproduction. ### Phi 3.5 Vision Accuracy: 69.3. Result ID: 69940cfd3a6c3148491ab686. [Reporting paper]() [Method origin paper]() Recorded evaluation configuration: {"# Vis tok.":null}. This record does not assert independent reproduction. ### GPT-4V (0409) Accuracy: 69.1. Result ID: 698b91007a967eddb76fa309. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Model Version":"0409"}. This record does not assert independent reproduction. ### GPT-4V Accuracy: 69.1. Result ID: 69936a983a6c31484915c851. [Reporting paper]() Recorded evaluation configuration: Not recorded. This record does not assert independent reproduction. ### LLaVA next 8B Accuracy: 63.8. Result ID: 69940cfd3a6c3148491ab688. [Reporting paper]() Recorded evaluation configuration: {"# Vis tok.":2880}. This record does not assert independent reproduction. ### Phantom-1.8B Accuracy: 63.1. Result ID: 698b91007a967eddb76fa30b. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"Size":"1.8B"}. This record does not assert independent reproduction. ### Mini-Gemini-HD 8B Accuracy: 62.6. Result ID: 69940cfd3a6c3148491ab689. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"# Vis tok.":2880}. This record does not assert independent reproduction. ### Vila 3B Accuracy: 55.2. Result ID: 69940cfd3a6c3148491ab685. [Reporting paper]() [Method origin paper]() [Implementation]() Recorded evaluation configuration: {"# Vis tok.":null}. This record does not assert independent reproduction. Page 1 of 1. This is an index page, not the full collection.