GOT-OCR 2.0
Unified end-to-end OCR model that reads documents like a vision-language model.
About
GOT-OCR 2.0 is StepFun's unified OCR model, built on the idea of OCR-2.0: a single end-to-end model that handles text recognition, layout, tables, charts, and handwriting instead of a pipeline of specialized detectors. Released in September 2024 under Apache 2.0, it hit number one on Hugging Face trending and passed a million downloads within months, with community ports to ONNX, llama.cpp, PaddleMIX, and ms-swift fine-tuning.
Unlike classic OCR engines that work word by word, GOT-OCR 2.0 processes whole document pages as images and emits structured text, which makes it strong on complex layouts, mathematical notation, and multilingual documents. It merged into the Hugging Face transformers library in early 2025, so it now runs through the standard vision-language model APIs.
Specs
Context window
Pricing
Free quota
API access
Weights
Released
Benchmarks
Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.
Use cases
Trust score — 8/10
Scored 8/10 (verified tier, metrics-v1). Evidence: Verified via vendor site/docs/GitHub and reputable coverage, wave 96; Trending score 76/100 from researcher assessment.