multimodal benchmark
OmniBench
A novel multimodal benchmark designed to evaluate large language models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. Comprises 1,142 question-answer pairs covering 8 task categories from basic perception to complex inference, with a unique constraint that accurate responses require integrated understanding of all three modalities.
Updated Aug 7, 2026
Published models1
Registry coverage1
MetricScore
EvidenceB
OmniBench leaderboard
Sorted by the source-provided rank. Higher score is better according to the registry.
| 1 | AC | 56.1% | 100.0% | 1 | C |
OmniBench highlights
The top published results on this benchmark's own scale.
What is OmniBench?
Definition and scoring fields from the benchmark registry.
A novel multimodal benchmark designed to evaluate large language models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. Comprises 1,142 question-answer pairs covering 8 task categories from basic perception to complex inference, with a unique constraint that accurate responses require integrated understanding of all three modalities.
Scores are shown in ratio. The current registry marks this benchmark as not independently verified with evidence level B.
- Family
- OmniBench
- Modality
- multimodal
- Primary category
- multimodal
- Score direction
- higher
- LLMBoard eligible
- No
- Evaluation key
- omnibench|llm-stats-current
Source-native results are preserved. Eligibility for the overall LLMBoard score is a separate policy decision.
FAQ
Common questions about OmniBench.
Which model scores highest on OmniBench?
Qwen2.5-Omni-7B is currently ranked first with 56.1%.
What does OmniBench measure?
A novel multimodal benchmark designed to evaluate large language models' ability to recognize, interpret, and reason across visual, acoustic, and textual inputs simultaneously. Comprises 1,142 question-answer pairs covering 8 task categories from basic perception to complex inference, with a unique constraint that accurate responses require integrated understanding of all three modalities.
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
1 unique published model results are currently shown.
Does this benchmark affect the overall score?
This benchmark is preserved as source-native evidence but is not eligible for the current overall score.