OpenVLM Leaderboard (Overall Avg) openvlm-overall Leaderboard
OpenCompass OpenVLM Leaderboard main board — average score across the full OpenVLM benchmark suite (MMBench, MMStar, MMMU, MathVista, OCRBench, AI2D, HallusionBench and more). The community standard aggregate ranking for vision-language models. · Metric: Avg Score (higher is better)
| # | Model | Avg Score | Paper |
|---|---|---|---|
| 1 | SenseNova-V6-5-Pro | 82.20 | link |
| 2 | CongRong-v2.0 | 80.70 | link |
| 3 | SenseNova-V6-Pro | 80.40 | link |
| 4 | Gemini-2.5-Pro | 80.10 | link |
| 5 | GPT-5-20250807 | 79.90 | link |
| 6 | JT-VL-Chat-V3.0 | 79.90 | link |
| 7 | InternVL3-78B | 79.10 | link |
| 8 | BlueLM-2.6-3B | 78.40 | link |
| 9 | GPT-5-mini-20250807 | 78.00 | link |
| 10 | InternVL3-38B | 77.80 | link |
| 11 | Step-1o | 77.70 | link |
| 12 | SenseNova | 77.40 | link |
| 13 | InternVL2.5-78B-MPO | 77.00 | link |
| 14 | GLM-4v-Plus-20250111 | 76.70 | link |
| 15 | Ovis2-34B | 76.50 | link |
| 16 | HunYuan-Standard-Vision | 76.30 | link |
| 17 | Qwen2.5-VL-72B | 76.10 | link |
| 18 | GPT-4.1-20250414 | 75.90 | link |
| 19 | TeleMM | 75.90 | link |
| 20 | R-4B | 75.50 | link |
| 21 | ChatGPT-4o-latest | 75.40 | link |
| 22 | GPT-4.5 | 75.30 | link |
| 23 | InternVL2.5-38B-MPO | 75.30 | link |
| 24 | BailingMM-Pro-0120 | 75.20 | link |
| 25 | InternVL3-14B | 75.20 | link |
| 26 | Qwen2.5-VL-32B | 74.80 | link |
| 27 | Qwen2-VL-72B | 74.80 | link |
| 28 | Qwen-VL-Max-0809 | 74.40 | link |
| 29 | Kimi-VL-A3B-Thinking-2506 | 74.30 | link |
| 30 | InternVL3-8B | 73.60 | link |