OpenVLM MMBench V1.1 CN - Spatial Relationship: leaderboard
OpenCompass OpenVLM MMBench V1.1 Chinese split accuracy for spatial relationship reasoning.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Estimated already saturated. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5 Mini (2025-08-07) | 94.7 | #172 |
| 2 | GPT-5 | 93.3 | #98 |
| 3 | GPT-4.5 | 90.7 | #274 |
| 4 | Gemini 2.5 Pro | 89.3 | #152 |
| 5 | InternVL3-78B | 84 | #352 |
| 6 | GPT-4o ChatGPT | 82.7 | |
| 7 | GPT-5 Nano | 81.3 | #425 |
| 8 | GPT-4.1 (2025-04-14) | 80 | #250 |
| 9 | Qwen 2 VL 72B | 80 | #460 |
| 10 | Gemini 1.5 Pro (002) | 78.7 | #349 |
| 11 | Claude 3.7 Sonnet | 78.7 | #249 |
| 12 | InternVL2.5-78B | 78.7 | #431 |
| 13 | Grok 2 (1212) | 77.3 | #496 |
| 14 | GPT-4.1 Mini | 73.3 | #355 |
| 15 | Claude 3.5 Sonnet (20241022) | 73.3 | #291 |
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-spatial-relationship · How It Works · Data refreshed daily, snapshot 2026-10-20.