Benchmark
MMBench
Vision benchmarks measure how well models understand images. Scores are percentages.
All vision benchmarksScores from LLM Stats data.
- Models measured
- 6
- Scores recorded
- 6
- Sources
- 1
- IQ weight
- 0.39
Score distribution
- 84.5%
- Mean
- 86.5%
- Median
- 2.9%
- Std dev
- 81.9%
- 10th percentile
- 88%
- 90th percentile
- 79.6%
- Min
This benchmark has a weight of 0.39 for chat models, measured over 6 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.
Leaderboard
| # | Model | Score |
|---|---|---|
| 1 | Qwen2.5 VL 72B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 88% |
| 2 | Phi 4 Multimodal Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.7% |
| 3 | Qwen2-VL-72B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 86.5% |
| 4 | Qwen2.5-VL 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 84.3% |
| 5 | Phi-3.5 Vision Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 81.9% |
| 6 | DeepSeek VL2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim | 79.6% |
Data credits: Data by LLM Stats