SmolDocling 256M Preview needs about 0.8 GB of memory at Q4_K_M (a 0.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 24 quants, from 0.1 GB (IQ1_S) to 0.2 GB (Q6_K).
SmolDocling 256M Preview – compact vision‑language model for document conversion
SmolDocling 256M Preview is a 0.163 billion‑parameter vision‑language model. It supports OCR, layout, code, formula, chart, table, figure, caption, list, and full‑page conversion. The model is fully compatible with DoclingDocuments and can be used with transformers, vllm, or onnx.
Good at
- OCR and text extraction from images
- Layout and bounding‑box preservation
- Code and formula recognition
Limits
- No vision support in the base GGUF files
- Context length limited to 8192 tokens
- License is cdla‑permissive‑2.0
Best for
- Document conversion to markdown or HTML
- Batch inference of single‑page images
- Integration with Docling workflows
Who it’s for. Developers and researchers needing efficient document conversion in a lightweight model. It is suitable for local inference on modest hardware and for integration into Docling pipelines.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ1_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ1_M | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ2_XXS | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ2_XS | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ2_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ2_M | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q2_K_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ3_XXS | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ3_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ3_XS | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q2_K | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q3_K_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ3_M | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ4_XS | 0.1 GB | 0.8 GB | Fits | Get file → | |
| IQ4_NL | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q4_0 | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q3_K_M | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q3_K_L | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q4_1 | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q4_K_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q4_K_M | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q5_K_S | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q5_K_M | 0.1 GB | 0.8 GB | Fits | Get file → | |
| Q6_K | 0.2 GB | 0.9 GB | Fits | Get file → |
For 8 GB VRAM, use IQ1_S or IQ1_M (0.1 GB file, ~0.8 GB memory). If 12 GB VRAM is available, Q4_K_M (0.13 GB file, ~0.8 GB memory) offers a good balance. For 16 GB VRAM, Q6_K (0.17 GB file, ~0.9 GB memory) provides higher quality.
What GPU do you need for SmolDocling 256M Preview?
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 Q6_K (0.2 GB file, about 0.9 GB in use).
Which SmolDocling 256M Preview quant should you choose?
Q4_K_M is the usual starting point for SmolDocling 256M Preview: it balances file size and quality well. Here it is a 0.1 GB file.
If memory is tight, the smallest file is IQ1_S at 0.1 GB, with the quality trade-offs that come with fewer bits. Use it only when memory is very tight. Expect a clear drop in quality, which is usually worse on small models and easier to live with on large ones.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about SmolDocling 256M Preview
How much VRAM does SmolDocling 256M Preview need?
About 0.8 GB at Q4_K_M with an 8K context window. That is the 0.1 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 SmolDocling 256M Preview on an 8 GB GPU?
Yes, with Q6_K: the 0.2 GB file needs about 0.9 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of SmolDocling 256M Preview GGUF?
It depends on the quant: IQ1_S is 0.1 GB, IQ1_M is 0.1 GB, IQ2_XXS is 0.1 GB, IQ2_XS is 0.1 GB, IQ2_S is 0.1 GB, IQ2_M is 0.1 GB and 18 more.
What context length does SmolDocling 256M Preview support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 0.2 GB at 8K and 0.2 GB at 8K.
Where can I download SmolDocling 256M Preview GGUF?
The files are on Hugging Face in the mradermacher/SmolDocling-256M-preview-i1-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 SmolDocling 256M Preview use?
Hugging Face lists the license as cdla-permissive-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 SmolDocling 256M Preview on 8 GB VRAM?
Yes, the IQ1_S and IQ1_M quantizations are 0.1 GB and require approximately 0.8 GB memory at 8 k context.
Does this model support image input?
Yes, it is a vision‑language model with OCR, layout, and figure recognition capabilities.
What license applies to this model?
The model is released under cdla‑permissive‑2.0.