OpenVLM MMBench V1.1 CN - RR: leaderboard
OpenCompass OpenVLM MMBench V1.1 Chinese split accuracy for relation-reasoning questions.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around January 2027. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | InternVL2.5-78B | 87.2 | #431 |
| 2 | InternVL3-78B | 86.8 | #352 |
| 3 | GPT-5 | 85.7 | #98 |
| 4 | InternVL3-38B | 85.3 | #405 |
| 5 | InternVL3-8B | 83.7 | #620 |
| 6 | InternVL3-14B | 83.7 | #505 |
| 7 | GPT-5 Mini (2025-08-07) | 83.7 | #172 |
| 8 | Gemini 2.5 Pro | 82.9 | #152 |
| 9 | GPT-4.1 (2025-04-14) | 81.8 | #250 |
| 10 | Qwen 2 VL 72B | 81.4 | #460 |
| 11 | GPT-4.5 | 81.4 | #274 |
| 12 | GPT-4o ChatGPT | 81 | |
| 13 | GPT-4.1 Mini | 79.5 | #355 |
| 14 | Gemini 1.5 Flash (002) | 79.5 | #583 |
| 15 | Qwen 2 VL 7B | 79.1 | #750 |
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-rr · How It Works · Data refreshed daily, snapshot 2026-10-20.