OpenVLM MMBench V1.1 CN - Image Topic: leaderboard
OpenCompass OpenVLM MMBench V1.1 Chinese split accuracy for image topic recognition.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | 98.9 | #152 |
| 2 | InternVL3-8B | 98.9 | #620 |
| 3 | GPT-5 Nano | 98.9 | #425 |
| 4 | InternVL3-38B | 98.9 | #405 |
| 5 | InternVL3-78B | 98.9 | #352 |
| 6 | InternVL3-14B | 98.9 | #505 |
| 7 | Qwen 2 VL 72B | 98.9 | #460 |
| 8 | GPT-5 Mini (2025-08-07) | 98.9 | #172 |
| 9 | Aquila-VL-2B | 98.9 | #1238 |
| 10 | Claude 3.5 Sonnet | 98.9 | #350 |
| 11 | Gemma 3 4B | 98.9 | #1089 |
| 12 | InternVL2.5-78B | 98.9 | #431 |
| 13 | Llama 3.2 90B Vision Instruct | 98.9 | #622 |
| 14 | GPT-4o ChatGPT | 98.9 | |
| 15 | GPT-5 | 97.8 | #98 |
No result here: #3 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=openvlm-mmbench-v1-1-cn-image-topic · How It Works · Data refreshed daily, snapshot 2026-10-20.