MicrosoftReleased Sep 2021MIT
TrOCR has no score on the current Vision Evals. On the earlier Vision Evals board (last updated July 9, 2026) it passed 26.2% of OCR tasks (#56 of 58).
It is an open-weight model from Microsoft under the MIT license, and free for commercial use.
TrOCR has not yet been evaluated on the current benchmark. The results below are from the legacy version of Vision Evals, our previous benchmark. See the current Vision Evals
| Category | Passed | Score |
|---|---|---|
| Text Recognition | 16 / 30 | 53.3% |
| Focused Scene OCR | 34 / 99 | 34.3% |
| License Plate Recognition | 10 / 30 | 33.3% |
| VQA & Extraction | 0 / 60 | 0% |
| Handwritten Math | 0 / 10 | 0% |
Scores based on a single evaluation run · Methodology
View all legacy Vision Evals results →TrOCR (Transformer-based Optical Character Recognition) is an end-to-end OCR model released in September 2021 by Microsoft Research. It departs from the traditional two-stage OCR pipeline — which typically combines a CNN-based feature extractor with an RNN-based sequence decoder — by using a pure Transformer architecture composed of a pretrained image Transformer encoder and a pretrained text Transformer decoder, an approach that later became standardized as the VisionEncoderDecoder pattern in Hugging Face Transformers.
TrOCR is released under MIT, a permissive license. The TrOCR license lets you use, modify, and sell work built on the model, with the copyright notice as the only real obligation and no requirement to open-source related code changes.
MIT grants no explicit patent license and disclaims all warranties. If patent exposure is a concern for your deployment, review it with counsel before launch.
Read the full MIT license ↗No commercial license is needed for TrOCR: permissive terms let you keep related code private while deploying commercially.
Do not hesitate to reach out with questions for your commercial project. Our team will help you start solving business problems on the first call. See Roboflow commercial licensing for the models included in each plan.
Talk to salesThis model is released under the MIT License, a short and permissive open-source license that allows commercial use, modification, and redistribution.
Yes. Under the terms of the MIT license, you can freely use this model for commercial purposes. You must retain the copyright notice and license text when redistributing.
License information is provided as a guide and is not legal advice.
Estimated cost per task vs. OCR score, for this model and others ranked near it. Upper-left is the sweet spot (high quality, low cost). Based on Vision Evals (legacy) results.
6 of 7 models plotted · 1 not yet evaluated
| Model | Score | Median tokens | Est. cost / task | Compare |
|---|---|---|---|---|
| Claude Sonnet 4.5 | 67.3% | 866 | $0.0039 | Compare |
| Gemma 3 4B | 64.2% | 314 | <$0.0001 | Compare |
| GPT-5.4 Nano | 62.5% | 294 | $0.0001 | Compare |
| Claude Haiku 4.5 | 61.6% | 861 | $0.0012 | Compare |
| Qwen3.6 Plus | 58.5% | 166 | $0.0001 | Compare |
| TrOCR(this model) | 26.2% | — | — | — |
| Kimi K2.5 | 19.6% | 706 | $0.0007 | Compare |
Other models worth comparing for similar use cases.