multimodal benchmark
MMBench Leaderboard
MMBench is a bilingual benchmark for assessing the multimodal capabilities of vision-language models through multiple-choice questions in English and Chinese.
Updated Sep 5, 2026
Models9
Model coverage9
MetricScore
EvidenceB
MMBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelST | Score91.80% | Percentile100.00% | Participants9 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score88.00% | Percentile87.50% | Participants9 | EvidenceC | Evaluated |
| Rank03 | ModelMI | Score86.70% | Percentile75.00% | Participants9 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score86.50% | Percentile62.50% | Participants9 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score84.30% | Percentile50.00% | Participants9 | EvidenceC | Evaluated |
| Rank06 | ModelMI | Score81.90% | Percentile37.50% | Participants9 | EvidenceC | Evaluated |
| Rank07 | ModelDE | Score80.30% | Percentile25.00% | Participants9 | EvidenceC | Evaluated |
| Rank08 | ModelDE | Score79.60% | Percentile12.50% | Participants9 | EvidenceC | Evaluated |
| Rank09 | ModelDE | Score69.20% | Percentile0.00% | Participants9 | EvidenceC | Evaluated |
MMBench Highlights
The leading models and scores on this benchmark.
MMBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for MMBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is MMBench?
What MMBench measures and how its scores work.
MMBench is a multimodal benchmark that evaluates vision-language models across diverse vision-language tasks using multiple-choice questions in English and Chinese.
MMBench measures vision-language model performance using the Score metric, reported as a ratio.
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- MMBench
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- overall
LLM Stats. Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about MMBench.
Which model scores highest on MMBench?
Step3-VL-10B is currently ranked first with 91.80%.
What are the top three models on MMBench?
The current leaders are Step3-VL-10B (91.80%), Qwen2.5 VL 72B Instruct (88.00%), and Phi-4-multimodal-instruct (86.70%).
Which MMBench model has the lowest official input price?
Qwen2.5 VL 7B Instruct has the lowest matched official input price at $0.35 input / $1.1 output per 1M tokens.
Which models are fastest among MMBench results?
The fastest matched records are Phi-4-multimodal-instruct (25.00 tok/s via DeepInfra).
Does the highest score result prove overall model quality?
No. This benchmark measures one defined capability or task. The overall LLMBoard score uses a separate aggregation across eligible benchmark evidence.