OpenVLM MMBench V1.1 CN - Nature Relation: leaderboard
OpenCompass OpenVLM MMBench V1.1 Chinese split accuracy for natural relation reasoning.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-4.5 | 96.7 | #274 |
| 2 | GPT-5 | 95.7 | #98 |
| 3 | GPT-4.1 (2025-04-14) | 95.7 | #250 |
| 4 | InternVL2.5-78B | 95.7 | #431 |
| 5 | Gemini 1.5 Flash (002) | 94.6 | #583 |
| 6 | InternVL3-78B | 94.6 | #352 |
| 7 | GPT-5 Mini (2025-08-07) | 94.6 | #172 |
| 8 | GPT-4o ChatGPT | 94.6 | |
| 9 | Gemini 2.5 Pro | 93.5 | #152 |
| 10 | Grok 2 (1212) | 93.5 | #496 |
| 11 | GPT-4.1 Mini | 91.3 | #355 |
| 12 | Claude 3.5 Sonnet (20241022) | 91.3 | #291 |
| 13 | Claude 3.7 Sonnet | 90.2 | #249 |
| 14 | Gemini 1.0 Pro | 90.2 | #933 |
| 15 | InternVL3-8B | 89.1 | #620 |
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-nature-relation · How It Works · Data refreshed daily, snapshot 2026-10-20.