What are the hardware requirements for LightOnOCR 3 4B?

LightOnOCR 3 4B requires approximately 9.1 GB of VRAM. Quantized variants may run on less VRAM; see the Quantized Models section on this page.

Who developed LightOnOCR 3 4B and how large is it?

LightOnOCR 3 4B is developed by lightonai, a model with 4b parameters. The model is published as open weights on Hugging Face and indexed on LLM Explorer with full benchmark history.

Where can I download or evaluate LightOnOCR 3 4B?

LightOnOCR 3 4B is hosted on Hugging Face and linked from this page. LLM Explorer also lists quantized variants and similar alternatives if available.

LightOnOCR 3 4B by lightonai — VRAM 9.1GB

Name: LightOnOCR 3 4B
Author: lightonai

LightOnOCR 3 4B is an open-source language model by lightonai. Features: 4b LLM, VRAM: 9.1GB, License: apache-2.0, LLM Explorer Score: 0.44.

Arxiv:2412.13663 Arxiv:2601.14251 Conversational Document-understanding Endpoints compatible Eval-results Forms Image-text-to-text Ocr Qwen3 5 Region:eu Safetensors Tables Vision Vision-language Visual-grounding

Model Card on HF 🤗: https://huggingface.co/lightonai/LightOnOCR-3-4B

LightOnOCR 3 4B Parameters and Internals

LLM Name	LightOnOCR 3 4B
Repository 🤗	https://huggingface.co/lightonai/LightOnOCR-3-4B
Model Size	4b
Required VRAM	9.1 GB
Updated	2026-10-11
Maintainer	lightonai
Model Type	qwen3_5
Model Files	9.1 GB
Model Architecture	Qwen3_5ForConditionalGeneration
License	apache-2.0
Model Max Length	262144
Transformers Version	5.5.4
Tokenizer Class	TokenizersBackend
Padding Token	<\|endoftext\|>
Errors	replace

Rank the LightOnOCR 3 4B capabilities

Have you tried this model? Rate its performance — this feedback helps the ML community find the right model for their needs.

Instruction Following and Task Automation
Factuality and Completeness of Knowledge
Censorship and Alignment
Data Analysis and Insight Generation
Text Generation
Text Summarization and Feature Extraction
Code Generation
Multi-Language Support and Translation

Best Alternatives to LightOnOCR 3 4B

Best Alternatives	Context / RAM	Downloads	Likes
AgentHorse 4B	256K / 8.4 GB	760	7
Qwen3.5 4B	0K / 9.3 GB	6575502	801
NuExtract3	0K / 0.2 GB	441570	324
FrogNano 4B 2609	0K / 9.3 GB	733	113
Agents A1 4B	0K / 9.1 GB	177578	72
Intern Decision 4B	0K / 0.1 GB	338	48
Tev1 4B Experimental	0K / 9.3 GB	1539	16
H2o Lightning 4B	0K / 9.1 GB	261	16
DeskForge Qwen3.5 4B	0K / 9.1 GB	1319	5
Qwen3.5 4B Base	0K / 9.3 GB	227814	79