image to text benchmark
OCRBench Leaderboard
OCRBench: Comprehensive evaluation benchmark for assessing Optical Character Recognition (OCR) capabilities in Large Multimodal Models across text recognition, scene text VQA, and document understanding tasks
Updated Aug 17, 2026
Models24
Model coverage24
MetricScore
EvidenceB
OCRBench Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelMA | Score92.3% | Percentile100.0% | Participants24 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score92.1% | Percentile95.7% | Participants24 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score92.0% | Percentile91.3% | Participants24 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score91.0% | Percentile87.0% | Participants24 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score90.3% | Percentile82.6% | Participants24 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score89.6% | Percentile78.3% | Participants24 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score89.5% | Percentile73.9% | Participants24 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score89.4% | Percentile69.6% | Participants24 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score89.4% | Percentile65.2% | Participants24 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score88.5% | Percentile60.9% | Participants24 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score88.1% | Percentile56.5% | Participants24 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score87.7% | Percentile52.2% | Participants24 | EvidenceC | Evaluated |
| Rank13 | ModelAC | Score87.5% | Percentile47.8% | Participants24 | EvidenceC | Evaluated |
| Rank14 | ModelAC | Score86.4% | Percentile43.5% | Participants24 | EvidenceC | Evaluated |
| Rank15 | ModelAC | Score85.5% | Percentile39.1% | Participants24 | EvidenceC | Evaluated |
| Rank16 | ModelMI | Score84.4% | Percentile34.8% | Participants24 | EvidenceC | Evaluated |
| Rank17 | ModelLA | Score84.2% | Percentile30.4% | Participants24 | EvidenceC | Evaluated |
| Rank18 | ModelAC | Score83.9% | Percentile26.1% | Participants24 | EvidenceC | Evaluated |
| Rank19 | ModelDE | Score83.4% | Percentile21.7% | Participants24 | EvidenceC | Evaluated |
| Rank20 | ModelAC | Score81.9% | Percentile17.4% | Participants24 | EvidenceC | Evaluated |
| Rank21 | ModelDE | Score81.1% | Percentile13.0% | Participants24 | EvidenceC | Evaluated |
| Rank22 | ModelDE | Score80.9% | Percentile8.7% | Participants24 | EvidenceC | Evaluated |
| Rank23 | ModelAC | Score80.8% | Percentile4.3% | Participants24 | EvidenceC | Evaluated |
| Rank24 | ModelCO | Score79.2% | Percentile0.0% | Participants24 | EvidenceC | Evaluated |
OCRBench Highlights
The leading models and scores on this benchmark.
OCRBench Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for OCRBench
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
Ranking basisThis ocrbench AI model leaderboard uses descending score in the benchmark's original unit. The leaderboard ranking keeps matched price and speed data separate from benchmark evidence.
- 01MAKimi K2.5Moonshot AI
- Score
- 92.3%
- Price
- $0.60 input / $3.0 output per 1M tokens
Strengths
- Ranks #1 of 24 compared models
- 100th percentile on this benchmark
- C evidence result
Considerations
- This result measures OCRBench, not total model capability
- 02ACQwen3.5-122B-A10BAlibaba Cloud / Qwen Team
- Score
- 92.1%
- Price
- $0.40 input / $3.2 output per 1M tokens
Strengths
- Ranks #2 of 24 compared models
- 96th percentile on this benchmark
- C evidence result
Considerations
- This result measures OCRBench, not total model capability
- 03ACQwen3 VL 235B A22B InstructAlibaba Cloud / Qwen Team
- Score
- 92.0%
Strengths
- Ranks #3 of 24 compared models
- 91th percentile on this benchmark
- C evidence result
Considerations
- This result measures OCRBench, not total model capability
- 04ACQwen3.5-35B-A3BAlibaba Cloud / Qwen Team
- Score
- 91.0%
- Price
- $0.25 input / $2.0 output per 1M tokens
Strengths
- Ranks #4 of 24 compared models
- 87th percentile on this benchmark
- C evidence result
Considerations
- This result measures OCRBench, not total model capability
- 05ACQwen3 VL 30B A3B InstructAlibaba Cloud / Qwen Team
- Score
- 90.3%
Strengths
- Ranks #5 of 24 compared models
- 83th percentile on this benchmark
- C evidence result
Considerations
- This result measures OCRBench, not total model capability
Selection summary
Best AI Models for OCRBench
Kimi K2.5 currently leads OCRBench with 92.3%. It is the top model on this specific benchmark, while the best LLM for the broader task should also be checked against other benchmarks, price and runtime.
Use this leaderboard with the supporting benchmark results and coverage details above. A leaderboard position summarizes the selected ranking signal; it does not replace workload-specific testing.
What is OCRBench?
What OCRBench measures and how its scores work.
OCRBench: Comprehensive evaluation benchmark for assessing Optical Character Recognition (OCR) capabilities in Large Multimodal Models across text recognition, scene text VQA, and document understanding tasks
Scores are shown in ratio. This benchmark is not independently verified and has an evidence level of B.
- Family
- OCRBench
- Modality
- multimodal
- Primary category
- image to text
- Score direction
- higher
- LLMBoard eligible
- Yes
- Evaluation key
- ocrbench|llm-stats-current
Benchmark scores retain their original unit. Overall score eligibility is shown separately.
FAQ
Common questions about OCRBench.
Which model scores highest on OCRBench?
Kimi K2.5 is currently ranked first with 92.3%.
What does OCRBench measure?
OCRBench: Comprehensive evaluation benchmark for assessing Optical Character Recognition (OCR) capabilities in Large Multimodal Models across text recognition, scene text VQA, and document understanding tasks
Is a higher score better?
Yes. Higher values rank better for this benchmark.
How many models are compared?
24 model results are currently shown.
Does this benchmark affect the overall score?
Yes. This benchmark can contribute to the current LLMBoard capability score.