AI benchmarks and leaderboards are different things
Learn the difference between a benchmark, an evaluation run and a leaderboard, and use each to answer the right question.
LLM Stats
MMBench
A bilingual benchmark for assessing multi-modal capabilities of vision-language models through multiple-choice questions in both English and Chinese, providing systematic evaluation across diverse vision-language tasks with robust metrics.
Evaluation context and evidence
- Source
- LLM Stats
- Publisher
- Not recorded
- Benchmark release
- Not recorded
- Collected
- 2026-08-17
- Modality
- multimodal
These are source-reported observations, not independent reruns by this site. Check the task version, evaluation protocol, unit and reporting date before comparing models. Results from different setups are not a universal ranking.
Original sources
Reported results 9
Shown in source observation order, not ranked. Dates retain the precision provided by the source.
| Model | Value | Unit | Protocol | Reported date and basis | Source |
|---|---|---|---|---|---|
| Step3-VL-10B | 0.918 | Not recorded | Not recorded | 2026-01-15 · model_announcement | LLM Stats |
| Qwen2.5 VL 72B Instruct | 0.88 | Not recorded | Not recorded | 2025-01-26 · model_announcement | LLM Stats |
| Phi-4-multimodal-instruct | 0.867 | Not recorded | Not recorded | 2025-02-01 · model_announcement | LLM Stats |
| Qwen2-VL-72B-Instruct | 0.865 | Not recorded | Not recorded | 2024-08-29 · model_announcement | LLM Stats |
| Qwen2.5 VL 7B Instruct | 0.843 | Not recorded | Not recorded | 2025-01-26 · model_announcement | LLM Stats |
| Phi-3.5-vision-instruct | 0.819 | Not recorded | Not recorded | 2024-08-23 · model_announcement | LLM Stats |
| DeepSeek VL2 Small | 0.803 | Not recorded | Not recorded | 2024-12-13 · model_announcement | LLM Stats |
| DeepSeek VL2 | 0.796 | Not recorded | Not recorded | 2024-12-13 · model_announcement | LLM Stats |
| DeepSeek VL2 Tiny | 0.692 | Not recorded | Not recorded | 2024-12-13 · model_announcement | LLM Stats |