Unlimited-OCR-2
baidumittextvision

Unlimited-OCR-2

Updated Nov 21, 2026
Parameters
3.3B
Context
131,072
License
mit
Q4 GGUF
2.02 GB
Updated
Nov 21, 2026

Q4 GGUF download: AtomicChat/Unlimited-OCR-2-GGUF - 2.02 GB on disk, read 2026-11-19. VRAM figures above are estimates; this is the actual file size.

Will it run on your hardware?

Pick your GPU memory - see which quantizations fit, and the cheapest card for the rest

Runs on your 24 GB - best at FP16
4 of 4 quantizations fit Unlimited-OCR-2 with real runtime overhead.
Q4_K_M
2 GB
Q5_K_M
2 GB
Q8_0
4 GB
FP16
7 GB
fits tight too big

Need an exact figure for your context length? Use the VRAM calculator.

Run it locally

Copy-paste - running in under a minute

vLLMOpenAI-compatible API
vllm serve baidu/Unlimited-OCR-2

New to this? Start with Ollama · serve to many users with vLLM.

Deep dive

Notes, sources, and the full write-up

Unlimited-OCR-2 is an open-weight vision-language model published by baidu on Hugging Face in October 2026, with 3.3B parameters under the mit license.

Among the 74 vision-language models in this directory it ranks 26th by monthly Hugging Face downloads, and 64 of the others are larger.

HardwareLink to this section

Unquantized FP16 weights take about 7 GB; 8-bit brings that to ~4 GB and 4-bit (Q4_K_M) to ~2 GB, before activations and any context or image buffers. A 12 GB consumer card holds it at 4-bit; FP16 needs a 24 GB card such as an RTX 4090. Check exact numbers for your card with the VRAM calculator.

Similar vision-language modelsLink to this section

PopularityLink to this section

Unlimited-OCR-2 was downloaded 612,067 times on Hugging Face in the last month (612,067 all-time) and has 717 likes.

Run it locallyLink to this section

Weights and usage instructions are on the baidu/Unlimited-OCR-2 model card; it loads with Hugging Face Transformers.

Local runtimes that fit this model:

  • vllm - high-throughput serving on NVIDIA GPUs
  • transformers - the reference Python library most model cards target

PaperLink to this section

From the community

What users are saying

“Would it be possible to provide an MLX-compatible version of Unlimited-OCR-2? I'd love to run this model locally on my Mac mini M4 Pro using MLX. Thanks for your great work!”

“We packaged Unlimited-OCR-2 as single CMF files that run with cortiq, a Rust inference engine with no Python: the DeepEncoder (SAM ViT-B and CLIP-L), the MoE decoder and DASH attention, on Vulkan/DX12/Metal or CPU. cortiq serve gives an OpenAI-compatible endpoint with image parts.”

Frequently asked

Quick answers to common questions

How much VRAM does Unlimited-OCR-2 need?

Unlimited-OCR-2 with 3.3B parameters needs approximately 2 GB at Q4_K_M quantization. Use our VRAM calculator for an exact estimate.

Is Unlimited-OCR-2 better than other baidu models?

Unlimited-OCR-2 has 3.3B parameters with 131,072 context - a strong choice for vision, long-context, edge.

What license is Unlimited-OCR-2 under?

Unlimited-OCR-2 is released under the mit license, making it suitable for most commercial and personal projects.

What hardware runs Unlimited-OCR-2 well?

With 3.3B parameters, Unlimited-OCR-2 requires adequate VRAM. High-end GPUs like the RTX 4090 (24GB), RTX 5090 (32GB), or Mac Studio with unified memory are good options. Check our hardware directory for specific recommendations.

What is the best quantization for Unlimited-OCR-2?

Q4_K_M is the recommended sweet spot - ~98% of FP16 quality at ~27% of the size. Q5_K_M (~2 GB) is an option if you have spare VRAM. Use our VRAM calculator to compare.

How long can Unlimited-OCR-2's context window handle?

Unlimited-OCR-2 supports a 131,072-token context window - enough for very long documents, codebases, or multi-turn conversations. Real-world usable context may vary by implementation.

What models compete with Unlimited-OCR-2?

Unlimited-OCR-2 competes with other models in its class. Browse our model directory for comparisons, benchmarks, and community reviews to find the best fit.

Can my GPU run Unlimited-OCR-2?

Fit and estimated speed on popular hardware

Nearby options

Similar models and compatible hardware by spec