Vision Understanding
Models for image analysis, visual question answering, and scene understanding
Last Updated
Nov 9, 2025
Total Votes
85,580
Total Models
10
Single Column
| Rank (UB) | Model | Score | Win% [?] | Votes | Organization | License | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| 🥇 | GPT-4 Vision | 94 | 79 | 101,946 | OpenAI | Proprietary | 🥇 GPT-4 VisionOpenAI Score94 Win%79 Votes101,946 | ||||||||
| 🥈 | Claude 3 Opus | 93 | 92 | 101,604 | Anthropic | Proprietary | 🥈 Claude 3 OpusAnthropic Score93 Win%92 Votes101,604 | ||||||||
| 🥉 | Gemini Ultra Vision | 93 | 83 | 97,055 | Proprietary | 🥉 Gemini Ultra VisionScore93 Win%83 Votes97,055 | |||||||||
| 4 | LLaVA 1.6 | 87 | 80 | 93,622 | Open Source | Proprietary | 4 LLaVA 1.6Open Source Score87 Win%80 Votes93,622 | ||||||||
| 5 | Qwen-VL-Plus | 86 | 87 | 91,152 | Alibaba | Proprietary | 5 Qwen-VL-PlusAlibaba Score86 Win%87 Votes91,152 | ||||||||
| 6 | CogVLM | 84 | 85 | 90,774 | Tsinghua | Proprietary | 6 CogVLMTsinghua Score84 Win%85 Votes90,774 | ||||||||
| 7 | BLIP-2 | 82 | 80 | 88,238 | Salesforce | Proprietary | 7 BLIP-2Salesforce Score82 Win%80 Votes88,238 | ||||||||
| 8 | InstructBLIP | 80 | 76 | 88,297 | Salesforce | Proprietary | 8 InstructBLIPSalesforce Score80 Win%76 Votes88,297 | ||||||||
| 9 | MiniGPT-4 | 79 | 68 | 80,112 | Open Source | Proprietary | 9 MiniGPT-4Open Source Score79 Win%68 Votes80,112 | ||||||||
| 10 | Flamingo | 77 | 75 | 82,389 | DeepMind | Proprietary | 10 FlamingoDeepMind Score77 Win%75 Votes82,389 | ||||||||