# OvisOCR3

> OvisOCR3 is a unified 0.8B multimodal model for OCR and document understanding. It integrates document parsing, general OCR, key information extraction, document figure captioning, and cross-page table reconstruction into a single compact model.

- **Organisation:** ATH-MaaS
- **Category:** OCR & Document AI, Vision-Language Models
- **Parameters:** 853M
- **Context length:** 262,144 tokens
- **Licence:** Apache 2.0 (Commercial use OK)
- **Released:** 2026-10-21
- **Hugging Face:** 437.7K downloads (30 days), 128 likes

## Estimated memory

| Quant | ≈ GB |
|---|---|
| Q4_K_M | 1.08 |
| Q8_0 | 1.51 |
| F16 | 2.41 |

## Runs with

- [llama.cpp](https://selfhost.directory/project/llamacpp)
- [SGLang](https://selfhost.directory/project/sglang)
- [Transformers](https://github.com/huggingface/transformers)
- [vLLM](https://selfhost.directory/project/vllm)
- [Jan](https://jan.ai) (community)
- [KoboldCpp](https://selfhost.directory/project/koboldcpp) (community)
- [LM Studio](https://lmstudio.ai) (community)
- [LocalAI](https://selfhost.directory/project/localai) (community)
- [Ollama](https://selfhost.directory/project/ollama) (community)
- [ONNX Runtime](https://onnxruntime.ai) (community)

## Benchmarks

- OmniDocBench (Real5): 93.71% (independent, rank 1/22)

## About

OvisOCR3 is a unified 0.8B multimodal model for OCR and document understanding. It integrates document parsing, general OCR, key information extraction, document figure captioning, and cross-page table reconstruction into a single compact model. These capabilities are selected through task-specific prompts, bringing recognition, extraction, and understanding together without requiring separate model deployments for each task.

Built on Qwen3.5-0.8B, OvisOCR3 is trained in four stages: continued pretraining, mid-training, on-policy distillation, and reinforcement learning. This training process builds a strong foundation in text recognition and document structure understanding, then refines these capabilities through domain-specific supervision and rewards.

OvisOCR3 achieves state-of-the-art performance among models of comparable size across all five domains, demonstrating that a single compact model can combine strong document parsing with a broad range of OCR and document-understanding capabilities.

The task prompts specify the output contracts. The examples below use these files directly.

Performance comparisons across five task domains are reported below. -- denotes an unreported value, not zero. Bold scores indicate the best result in each column.

## Alternatives

- [datalab-to/surya-ocr-2](https://selfhost.directory/models/datalab-to/surya-ocr-2) (686M, OpenRAIL, ≈2.03 GB): It works on a range of documents (see usage and benchmarks).
- [datalab-to/surya-ocr-2.5](https://selfhost.directory/models/datalab-to/surya-ocr-2.5) (689M, OpenRAIL, ≈2.04 GB): We also ship smaller models for line-level text detection and ocr error detection. It works on a range of documents (see usage and benchmarks).
- [lightonai/LightOnOCR-2-1B](https://selfhost.directory/models/lightonai/lightonocr-2-1b) (1B, Apache 2.0, ≈2.75 GB): Best OCR model . LightOnOCR-2-1B is LightOn's flagship OCR model, refined with RLVR training for maximum accuracy. We recommend this variant for most OCR tasks.
- [lightonai/LightOnOCR-3-0.8B](https://selfhost.directory/models/lightonai/lightonocr-3-0.8b) (853M, Apache 2.0, ≈2.41 GB): Smallest model. LightOnOCR-3-0.8B is the lightest model of LightOn's LightOnOCR-3 family, with the same layout and visual understanding features.
- [ATH-MaaS/OvisOCR2](https://selfhost.directory/models/ath-maas/ovisocr2) (853M, Apache 2.0, ≈2.41 GB): We are pleased to announce the release of OvisOCR2, a compact 0.8B end-to-end model for page-level document parsing. Given a document page image, OvisOCR2 generates a Markdown representation in natural reading order, covering text, formulas, tables, and visual regions.
- [PaddlePaddle/PaddleOCR-VL-1.6](https://selfhost.directory/models/paddlepaddle/paddleocr-vl-1.6) (959M, Apache 2.0, ≈2.64 GB): PaddleOCR-VL-1.6: Expanding the Frontier of Document Parsing with Under-Optimized Region Refinement and Progressive Post-Training

More: https://selfhost.directory/models/ath-maas/ovisocr3/alternatives

## Runs well on

- [GeForce RTX 4060](https://selfhost.directory/models/ath-maas/ovisocr3/on/rtx-4060): F16
- [Apple M4 (16 GB)](https://selfhost.directory/models/ath-maas/ovisocr3/on/apple-m4-16gb): F16
- [Radeon RX 6700 XT](https://selfhost.directory/models/ath-maas/ovisocr3/on/rx-6700-xt): F16
- [GeForce RTX 3060 12GB](https://selfhost.directory/models/ath-maas/ovisocr3/on/rtx-3060): F16
- [Arc B580](https://selfhost.directory/models/ath-maas/ovisocr3/on/arc-b580): F16
- [Tesla T4](https://selfhost.directory/models/ath-maas/ovisocr3/on/tesla-t4): F16

## Similar size

- [PaddlePaddle/PaddleOCR-VL-1.6](https://selfhost.directory/models/paddlepaddle/paddleocr-vl-1.6) (959M)
- [datalab-to/surya-ocr-2.5](https://selfhost.directory/models/datalab-to/surya-ocr-2.5) (689M)
- [lightonai/LightOnOCR-2-1B](https://selfhost.directory/models/lightonai/lightonocr-2-1b) (1B)
- [PaddlePaddle/PaddleOCR-VL](https://selfhost.directory/models/paddlepaddle/paddleocr-vl) (959M)
- [PaddlePaddle/PaddleOCR-VL-1.5](https://selfhost.directory/models/paddlepaddle/paddleocr-vl-1.5) (959M)
- [stepfun-ai/GOT-OCR2_0](https://selfhost.directory/models/stepfun-ai/got-ocr2_0) (716M)

## FAQ

**How much VRAM does OvisOCR3 need?**
About 2.4 GB at FP16 / BF16.

**Can OvisOCR3 run on the RTX 4060?**
Yes. It fits at FP16 / BF16 (about 2.2 GB). It fits comfortably on the Apple M4 (16 GB), Radeon RX 6700 XT, RTX 3060 12GB.

**Can I use OvisOCR3 commercially?**
Yes. It is released under Apache 2.0. Permissive: use, modify and sell freely; keep the notice and licence text.

**How do I run OvisOCR3 with Ollama or llama.cpp?**
With Ollama: ollama run hf.co/mradermacher/OvisOCR3-GGUF:Q4_K_M. With llama.cpp: llama-server -hf mradermacher/OvisOCR3-GGUF:Q4_K_M, which starts an OpenAI-compatible server.

## Links

- huggingface: https://huggingface.co/ATH-MaaS/OvisOCR3
- paper: https://arxiv.org/abs/2610.33066
- demo: https://huggingface.co/spaces/ATH-MaaS/OvisOCR3

---
Source: https://selfhost.directory/models/ath-maas/ovisocr3