image to text benchmark
OCRBench-V2 (en) Leaderboard
OCRBench-V2 (en) is a benchmark for evaluating Large Multimodal Models on visual text localization and reasoning with English text content.
Updated Sep 5, 2026
Models14
Model coverage14
MetricScore
EvidenceB
OCRBench-V2 (en) Ranking
Higher score ranks better on this benchmark.
Sort by
Rank | Model | Score | Percentile | Participants | Evidence | Evaluated |
|---|
| Rank01 | ModelAC | Score68.40% | Percentile100.00% | Participants14 | EvidenceC | Evaluated |
| Rank02 | ModelAC | Score67.40% | Percentile92.31% | Participants14 | EvidenceC | Evaluated |
| Rank03 | ModelAC | Score67.10% | Percentile84.62% | Participants14 | EvidenceC | Evaluated |
| Rank04 | ModelAC | Score66.80% | Percentile76.92% | Participants14 | EvidenceC | Evaluated |
| Rank05 | ModelAC | Score65.40% | Percentile69.23% | Participants14 | EvidenceC | Evaluated |
| Rank06 | ModelAC | Score63.90% | Percentile61.54% | Participants14 | EvidenceC | Evaluated |
| Rank07 | ModelAC | Score63.70% | Percentile53.85% | Participants14 | EvidenceC | Evaluated |
| Rank08 | ModelAC | Score63.20% | Percentile46.15% | Participants14 | EvidenceC | Evaluated |
| Rank09 | ModelAC | Score62.60% | Percentile38.46% | Participants14 | EvidenceC | Evaluated |
| Rank10 | ModelAC | Score61.80% | Percentile30.77% | Participants14 | EvidenceC | Evaluated |
| Rank11 | ModelAC | Score61.50% | Percentile23.08% | Participants14 | EvidenceC | Evaluated |
| Rank12 | ModelAC | Score57.20% | Percentile15.38% | Participants14 | EvidenceC | Evaluated |
| Rank13 | ModelLA | Score47.50% | Percentile7.69% | Participants14 | EvidenceC | Evaluated |
| Rank14 | ModelCO | Score36.70% | Percentile0.00% | Participants14 | EvidenceC | Evaluated |
OCRBench-V2 (en) Highlights
The leading models and scores on this benchmark.
OCRBench-V2 (en) Score Distribution
A closer view of the leading scores on this benchmark.
The Top AI Models for OCRBench-V2 (en)
The first five results on this benchmark, with official price and output speed added where the model identity can be matched.
What is OCRBench-V2 (en)?
What OCRBench-V2 (en) measures and how its scores work.
OCRBench-V2 (en) is the English subset of OCRBench v2 for evaluating Large Multimodal Models.
It measures visual text localization and reasoning with English text content 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
- OCRBench-V2 (en)
- Modality
- multimodal
- Primary category
- image to text
- 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 OCRBench-V2 (en).
Which model scores highest on OCRBench-V2 (en)?
Qwen3 VL 32B Thinking is currently ranked first with 68.40%.
What are the top three models on OCRBench-V2 (en)?
The current leaders are Qwen3 VL 32B Thinking (68.40%), Qwen3 VL 32B Instruct (67.40%), and Qwen3 VL 235B A22B Instruct (67.10%).
Which OCRBench-V2 (en) model has the lowest official input price?
No matched official input price is currently available.
Which models are fastest among OCRBench-V2 (en) results?
The fastest matched records are Qwen3 VL 4B Thinking (8.77 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.