multimodal benchmark
OmniDocBench Leaderboard
OmniDocBench evaluates multimodal models on document understanding tasks such as OCR, layout parsing, and structured document comprehension.
Updated Sep 5, 2026
Models2
Model coverage2
MetricScore
EvidenceB
OmniDocBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelMA | Score91.10% | Percentile100.00% | Participants2 | EvidenceC | Evaluated |
| Rank02 | ModelXI | Score87.20% | Percentile0.00% | Participants2 | EvidenceC | Evaluated |
OmniDocBench Highlights
The leading models and scores on this benchmark.
OmniDocBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for OmniDocBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is OmniDocBench?
What OmniDocBench measures and how its scores work.
OmniDocBench is a multimodal benchmark for evaluating models on document understanding tasks.
It measures performance on OCR, layout parsing, and structured document comprehension 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
- OmniDocBench
- 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 OmniDocBench.
Which model scores highest on OmniDocBench?
Kimi K3 is currently ranked first with 91.10%.
What are the top three models on OmniDocBench?
The current leaders are Kimi K3 (91.10%) and MiMo-V2.5 (87.20%).
Which OmniDocBench model has the lowest official input price?
MiMo-V2.5 has the lowest matched official input price at $0.14 input / $0.28 output per 1M tokens.
Which models are fastest among OmniDocBench results?
The fastest matched records are MiMo-V2.5 (5.39 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.