HU

HunyuanOCR

HunyuanocrGGUF by ggml-orgUpdated 2026-05-21
Download on Hugging Face
Parameters539M
Context32K
Layers24
Architecturehunyuan_vl
Smallest quant0.6 GB
Quick answer

HunyuanOCR needs about 1.9 GB of memory at Q8_0 (a 0.6 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 2 quants, from 0.6 GB (Q8_0) to 1.1 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 3.2 GB.

HunyuanOCR – lightweight OCR‑specialized vision‑language model

HunyuanOCR is a lightweight VLM focused on OCR tasks. It supports document parsing, text spotting, information extraction, and text‑image translation. The model is available in GGUF format for llama.cpp deployment.

Good at

  • OCR on documents
  • Text spotting in images
  • Information extraction from visual text

Limits

  • No vision support beyond OCR
  • Large memory requirement for full context
  • License not specified

Best for

  • OCR on laptops
  • Document parsing in low‑resource settings
  • Text extraction from scanned images

Who it’s for. Developers needing an OCR‑centric VLM that can run on consumer hardware. The model is suitable for applications that require end‑to‑end OCR without heavy GPU resources.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_00.6 GB 1.9 GB
Fits Get file →
BF161.1 GB 2.4 GB
Fits Get file →

For 8 GB VRAM, use the Q8_0 quant (0.58 GB, ~1.9 GB memory at 8k context). For higher precision on 16 GB VRAM, BF16 (1.08 GB, ~2.4 GB memory at 8k context) is recommended. Memory estimates are approximate.

What GPU do you need for HunyuanOCR?

These figures add the file size, an fp16 KV cache at an 8K context window (worked out from this model’s own header) and about 0.5 GB for the runtime. They are estimates, so leave a little headroom, and expect to use less context or a smaller quant if you are right at the limit.

  • 4 GB VRAM: the best quant that fits is BF16 (1.1 GB file, about 2.4 GB in use).

Which HunyuanOCR quant should you choose?

Q8_0 (0.6 GB) is the middle option in this repository. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturehunyuan_vl
Parameters539M
Layers24
Trained context32,768 tokens
KV cache (fp16)96 KB per token

Source

Quantized byggml-org
Licenseother
Downloads (HF)6,325
Last updated2026-05-21

Frequently asked questions about HunyuanOCR

How much VRAM does HunyuanOCR need?

About 1.9 GB at Q8_0 with an 8K context window. That is the 0.6 GB file, plus the KV cache and roughly 0.5 GB of runtime overhead. Smaller quants need less and longer context needs more.

Can I run HunyuanOCR on an 8 GB GPU?

Yes, with BF16: the 1.1 GB file needs about 2.4 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of HunyuanOCR GGUF?

It depends on the quant: Q8_0 is 0.6 GB, BF16 is 1.1 GB.

What context length does HunyuanOCR support?

The GGUF header lists a trained context length of 32,768 tokens (32K). The KV cache grows with the context you use: about 0.8 GB at 8K and 3.2 GB at 32K.

Where can I download HunyuanOCR GGUF?

The files are on Hugging Face in the ggml-org/HunyuanOCR-GGUF repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does HunyuanOCR use?

Hugging Face lists the license as other. Licenses can differ between the quantized files and the original model, so check the repository and the base model before commercial use.

Can I run HunyuanOCR on 8 GB VRAM?

Yes, the Q8_0 quant is 0.58 GB and uses about 1.9 GB of VRAM at 8k context, fitting within 8 GB.

Does HunyuanOCR support multi‑image text‑centric QA?

Yes, the model was trained with data for multi‑image text‑centric QA tasks.

Is there a license restriction for using HunyuanOCR?

The license is listed as "other"; refer to the original model card for details.

Other GGUFs of HunyuanOCR

Similar models

Browse all →

More from ggml-org

All by ggml-org →

More Hunyuanocr models

All Hunyuanocr →

Other tiny models

All tiny models →