OpenVLM MMBench V1.1 CN - Celebrity Recognition: leaderboard
OpenCompass OpenVLM MMBench V1.1 Chinese split accuracy for celebrity recognition.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Not forecast. 284 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 2 VL 72B | 95.2 | #460 |
| 2 | Gemini 2.5 Pro | 93.2 | #152 |
| 3 | Claude 3.5 Sonnet (20241022) | 91.2 | #291 |
| 4 | Gemini 1.5 Pro (002) | 91.2 | #349 |
| 5 | InternVL3-78B | 91.2 | #352 |
| 6 | Qwen 2.5 VL 7B | 91.2 | #643 |
| 7 | Gemini 1.5 Flash (002) | 90.5 | #583 |
| 8 | Gemma 3 27B | 89.8 | #607 |
| 9 | Qwen 2 VL 7B | 89.1 | #750 |
| 10 | InternVL2.5-78B | 89.1 | #431 |
| 11 | Grok 2 (1212) | 89.1 | #496 |
| 12 | Llama 3.2 90B Vision Instruct | 89.1 | #622 |
| 13 | InternVL3-38B | 87.1 | #405 |
| 14 | Gemini 1.5 Pro | 87.1 | #445 |
| 15 | Gemini 1.5 Flash | 85.7 | #618 |
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-celebrity-recognition · How It Works · Data refreshed daily, snapshot 2026-10-20.