Model card
DeepSeek OCR.
DeepSeekopen-sourceVision-Language OCR Model
DeepSeek's OCR model for document understanding.
§ 02 · Benchmarks
Every benchmark DeepSeek OCR has a recorded score for.
| # | Benchmark | Area · Task | Metric | Value | Rank | Date | Source |
|---|---|---|---|---|---|---|---|
| 01 | OmniDocBench | Vision & Documents · Document Parsing | composite | 86.5% | #24 | 2026-04-20 | source ↗ |
| 02 | olmOCR-Bench | Vision & Documents · Document Parsing | pass-rate | 75.7% | #24 | 2026-04-20 | source ↗ |
| 03 | olmOCR-Bench | — | pass-rate | 75.7% | #24 | 2025-10-21 | source ↗ |
Rank column shows this model’s position vs all other models scored on the same benchmark + metric (competitors after the slash). #1 in red means current SOTA. Sorted by rank, then newest result.
§ 03 · Strengths by area
Where DeepSeek OCR actually performs.
§ 06 · Sources & freshness
Where these numbers come from.
codesota-api
2
results
paperswithcode-public-api
1
result
1 of 3 rows marked verified. · first result 2025-10-21, latest 2026-04-20.