LI

lightonai/LightOnOCR-3-0.8B

LightonocrGGUF by tomngdevUpdated 2026-10-15
Download on Hugging Face
Parameters752M
Context256K
Layers24
Architectureqwen35
Smallest quant0.5 GB
Quick answer

lightonai/LightOnOCR-3-0.8B needs about 1.4 GB of memory at Q4_K_M (a 0.5 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 6 quants, from 0.5 GB (Q4_K_M) to 3.0 GB (F32). Longer context needs more: at 32K tokens the KV cache alone adds about 1.6 GB.

LightOnOCR‑3‑0.8B: lightweight OCR with vision and bounding‑box output

LightOnOCR‑3‑0.8B is a 0.8B‑parameter OCR model based on the Qwen3.5 vision‑language architecture. It provides transcription, grounding with bounding boxes, image descriptions, and chart data extraction. The model is released under Apache‑2.0 for research and commercial use.

Good at

  • OCR transcription
  • Bounding‑box grounding
  • Chart data extraction

Limits

  • No vision support beyond OCR
  • Requires 8k context for estimates
  • No multi‑expert architecture

Best for

  • Document OCR pipelines
  • Receipt and form processing
  • Table extraction

Who it’s for. Developers needing a small, fast OCR model that outputs text, bounding boxes, and visual descriptions. Suitable for integration into document‑understanding workflows without complex pipelines.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M0.5 GB 1.4 GB
Fits Get file →
Q6_K0.6 GB 1.5 GB
Fits Get file →
Q8_00.8 GB 1.7 GB
Fits Get file →
BF161.5 GB 2.4 GB
Fits Get file →
F161.5 GB (3 parts) 2.4 GB
Fits Get file →
F323.0 GB (6 parts) 3.9 GB
Fits Get file →

For 8 GB VRAM, use Q4_K_M (0.53 GB, 1.4 GB memory). For 12 GB VRAM, Q6_K (0.63 GB, 1.5 GB memory) or Q8_0 (0.81 GB, 1.7 GB memory). Higher precision quants require more VRAM.

What GPU do you need for lightonai/LightOnOCR-3-0.8B?

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 F32 (3.0 GB file, about 3.9 GB in use).

Which lightonai/LightOnOCR-3-0.8B quant should you choose?

Q4_K_M is the usual starting point for lightonai/LightOnOCR-3-0.8B: it balances file size and quality well. Here it is a 0.5 GB file.

The Q8_0 file is 0.8 GB, about 1.5 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

Architectureqwen35
Parameters752M
Layers24
Trained context262,144 tokens
KV cache (fp16)48 KB per token

Source

Quantized bytomngdev
Licenseapache-2.0
Downloads (HF)579
Last updated2026-10-15

Frequently asked questions about lightonai/LightOnOCR-3-0.8B

How much VRAM does lightonai/LightOnOCR-3-0.8B need?

About 1.4 GB at Q4_K_M with an 8K context window. That is the 0.5 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 lightonai/LightOnOCR-3-0.8B on an 8 GB GPU?

Yes, with F32: the 3.0 GB file needs about 3.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of lightonai/LightOnOCR-3-0.8B GGUF?

It depends on the quant: Q4_K_M is 0.5 GB, Q6_K is 0.6 GB, Q8_0 is 0.8 GB, BF16 is 1.5 GB, F16 is 1.5 GB, F32 is 3.0 GB.

What context length does lightonai/LightOnOCR-3-0.8B support?

The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 0.4 GB at 8K and 1.6 GB at 32K.

Where can I download lightonai/LightOnOCR-3-0.8B GGUF?

The files are on Hugging Face in the tomngdev/LightOnOCR-3-0.8B-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 lightonai/LightOnOCR-3-0.8B use?

Hugging Face lists the license as apache-2.0. 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 LightOnOCR‑3‑0.8B on 8 GB VRAM?

Yes, the Q4_K_M quant uses 1.4 GB memory at 8k context, fitting within 8 GB VRAM.

Does the model support multi‑expert inference?

No, LightOnOCR‑3‑0.8B is not a mixture of experts.

What license applies to this model?

The model is released under the Apache‑2.0 license for research and commercial use.

Other GGUFs of LightOnOCR-3-0.8B

Similar models

Browse all →

More from tomngdev

All by tomngdev →

More Lightonocr models

All Lightonocr →

Other tiny models

All tiny models →