JI

jina ocr v1

JinaGGUF by ahmadwUpdated 2026-10-12
Download on Hugging Face
Parameters2.93B
Context—
Layers—
Architecturedeepseek2-ocr
Smallest quant2.0 GB
Quick answer

jina ocr v1 needs about 2.7 GB of memory at Q4_K_M (a 2.0 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 3 quants, from 2.0 GB (Q4_K_M) to 5.9 GB (F16).

Jina OCR v1 – efficient document parsing with speculative decoding

Jina OCR v1 is an end‑to‑end document parsing model built on DeepSeek‑OCR. It uses a DeepEncoder vision tower that represents a 1024×1024 page in 256 visual tokens and a 3B‑parameter mixture‑of‑experts decoder. The model scores 91.14 on OmniDocBench v1.6 and 83.4 on olmOCR‑Bench.

Good at

  • High OCR accuracy
  • Fast decoding with FastMTP
  • Efficient serving on CPU and Apple Silicon

Limits

  • No vision support in llama.cpp without patch
  • Large VRAM required for higher context
  • License is cc‑by‑nc‑4.0

Best for

  • Batch OCR of scanned documents
  • Automated form extraction
  • Rapid prototype OCR services

Who it’s for. Developers needing a lightweight OCR model for local deployment. The model is suitable for applications that require high throughput and can tolerate a non‑commercial license. It works best on systems with sufficient VRAM and CPU resources.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M2.0 GB 2.7 GB
Fits Get file →
Q8_03.1 GB 4.0 GB
Fits Get file →
F165.9 GB 7.1 GB
Fits Get file →

For 8 GB VRAM, use the Q4_K_M quant (1.95 GB file, ~2.7 GB memory at 8k context). If 12 GB VRAM is available, Q8_0 (3.13 GB file, ~4 GB memory) offers higher precision. For maximum quality and 16 GB VRAM, F16 (5.88 GB file, ~7.1 GB memory) is recommended.

What GPU do you need for jina ocr v1?

These figures add the file size, an fp16 KV cache at an 8K context window 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 Q8_0 (3.1 GB file, about 4.0 GB in use).
  • 6 GB VRAM: the best quant that fits is Q8_0 (3.1 GB file, about 4.0 GB in use).
  • 8 GB VRAM: the best quant that fits is F16 (5.9 GB file, about 7.1 GB in use).

Which jina ocr v1 quant should you choose?

Q4_K_M is the usual starting point for jina ocr v1: it balances file size and quality well. Here it is a 2.0 GB file.

The Q8_0 file is 3.1 GB, about 1.6 times the size of Q4_K_M. It keeps the weights closer to the original, which only matters if you have the memory to spare.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architecturedeepseek2-ocr
Parameters2.93B

Source

Quantized byahmadw
FamilyJina
Downloads (HF)2,105
Last updated2026-10-12

Frequently asked questions about jina ocr v1

How much VRAM does jina ocr v1 need?

About 2.7 GB at Q4_K_M with an 8K context window. That is the 2.0 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 jina ocr v1 on an 8 GB GPU?

Yes, with F16: the 5.9 GB file needs about 7.1 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of jina ocr v1 GGUF?

It depends on the quant: Q4_K_M is 2.0 GB, Q8_0 is 3.1 GB, F16 is 5.9 GB.

Where can I download jina ocr v1 GGUF?

The files are on Hugging Face in the ahmadw/jina-ocr-v1-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 jina ocr v1 use?

Hugging Face lists the license as cc-by-nc-4.0. 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 Jina OCR v1 on 8 GB VRAM?

Yes, the Q4_K_M quant file is 1.95 GB and requires about 2.7 GB of VRAM at 8k context.

What patch is needed for llama.cpp?

A one‑line patch to remove the rope_theta assert is required to run the GGUF files.

Is the model commercial?

The model is licensed cc‑by‑nc‑4.0, so it is non‑commercial and requires attribution.

Other GGUFs of jina-ocr-v1

Similar models

Browse all →

More from ahmadw

All by ahmadw →

More Jina models

All Jina →

Other tiny models

All tiny models →