OCR models / MonkeyOCR team
MonkeyOCR v2 (B)
MonkeyOCR team · released 2026-07-11 · data checked 2026-11-06
MonkeyOCR v2 (B) is a 880M-parameter OCR and document-parsing model from MonkeyOCR team, released 2026-07-11 under the Apache-2.0 license. It scores 91.80 on OmniDocBench v1.6 (#23 of 33). It needs about 2.8 GB of memory at 16-bit.
OmniDocBench v1.6
91.80
#23 of 33 models
olmOCR-Bench
—
Not reported
Parameters
880M
OCR specialist
Memory to run
~2.8 GB
16-bit weights plus about 1 GB
Benchmarks
OmniDocBench v1.6 Leaderboard91.80
The leaderboard entry "MonkeyOCR V2" does not say which size was tested. Best: OXR 1.0 97.03
The tick marks the best score on each benchmark. See how to read the ranking.
Specifications
- Lab
- MonkeyOCR team
- Type
- OCR / document parsing specialist
- Total parameters
- 880M
- License
- Apache-2.0
- Memory to run
- ~2.8 GB at 16-bit
- Released
- 2026-07-11
Closest alternatives
- Infinity-Parser2 Flash (2.2B, 91.98 on OmniDocBench, Apache-2.0)
- jina-ocr-v1 (3.4B, 91.14 on OmniDocBench, CC-BY-NC-4.0)
- dots.ocr (3B, 90.77 on OmniDocBench, MIT)
- MinerU2.5 (2509) (1.2B, 93.04 on OmniDocBench, AGPL-3.0)
Sources
Something wrong here? Report it.