SmolVLM2 2.2B Instruct needs about 3.2 GB of memory at Q4_K_M (a 1.1 GB file) with an 8K context window, so it fits on a 4 GB GPU. This repository has 13 quants, from 0.7 GB (Q2_K) to 3.6 GB (F16).
SmolVLM2 2.2B Instruct – lightweight multimodal model
SmolVLM2 2.2B Instruct is a compact multimodal model that handles video, image, and text inputs. It generates text responses to queries about media content. The model is based on the Smolvlm family and uses a llama architecture.
Good at
- Vision‑text inference
- Video captioning
- Visual question answering
Limits
- No image or video generation
- Context size limited to 8192 tokens
- Requires 5.2 GB GPU RAM for video inference
Best for
- On‑device video analysis
- Multimodal chatbots
- Captioning short clips
Who it’s for. Developers building lightweight multimodal applications on devices with limited GPU memory. Researchers needing a small vision‑language model for rapid prototyping.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q2_K | 0.7 GB | 2.8 GB | Fits | Get file → | |
| Q3_K_S | 0.8 GB | 2.9 GB | Fits | Get file → | |
| Q3_K_M | 0.9 GB | 3.0 GB | Fits | Get file → | |
| Q3_K_L | 1.0 GB | 3.1 GB | Fits | Get file → | |
| Q4_0 | 1.0 GB | 3.2 GB | Fits | Get file → | |
| Q4_K_S | 1.1 GB | 3.2 GB | Fits | Get file → | |
| Q4_K_M | 1.1 GB | 3.2 GB | Fits | Get file → | |
| Q5_0 | 1.3 GB | 3.4 GB | Fits | Get file → | |
| Q5_K_S | 1.3 GB | 3.4 GB | Fits | Get file → | |
| Q5_K_M | 1.3 GB | 3.4 GB | Fits | Get file → | |
| Q6_K | 1.5 GB | 3.6 GB | Fits | Get file → | |
| Q8_0 | 1.9 GB | 4.0 GB | Fits | Get file → | |
| F16 | 3.6 GB | 5.7 GB | Fits | Get file → |
For 8 GB VRAM use Q4_K_M (est. 3.2 GB). For 12 GB VRAM use Q5_K_M (est. 3.4 GB). For 16 GB VRAM use Q6_K (est. 3.6 GB).
What GPU do you need for SmolVLM2 2.2B Instruct?
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 Q8_0 (1.9 GB file, about 4.0 GB in use).
- 6 GB VRAM: the best quant that fits is F16 (3.6 GB file, about 5.7 GB in use).
Which SmolVLM2 2.2B Instruct quant should you choose?
Q4_K_M is the usual starting point for SmolVLM2 2.2B Instruct: it balances file size and quality well. Here it is a 1.1 GB file.
The Q8_0 file is 1.9 GB, about 1.7 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.
If memory is tight, the smallest file is Q2_K at 0.7 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 SmolVLM2 2.2B Instruct
How much VRAM does SmolVLM2 2.2B Instruct need?
About 3.2 GB at Q4_K_M with an 8K context window. That is the 1.1 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 SmolVLM2 2.2B Instruct on an 8 GB GPU?
Yes, with F16: the 3.6 GB file needs about 5.7 GB at 8K context. Larger quants in this repository do not fit in 8 GB.
What is the file size of SmolVLM2 2.2B Instruct GGUF?
It depends on the quant: Q2_K is 0.7 GB, Q3_K_S is 0.8 GB, Q3_K_M is 0.9 GB, Q3_K_L is 1.0 GB, Q4_0 is 1.0 GB, Q4_K_S is 1.1 GB and 7 more.
What context length does SmolVLM2 2.2B Instruct support?
The GGUF header lists a trained context length of 8,192 tokens (8K). The KV cache grows with the context you use: about 1.6 GB at 8K and 1.6 GB at 8K.
Where can I download SmolVLM2 2.2B Instruct GGUF?
The files are on Hugging Face in the second-state/SmolVLM2-2.2B-Instruct-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 SmolVLM2 2.2B Instruct 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 SmolVLM2 on 8 GB VRAM?
Yes, the Q4_K_M quantized model uses about 3.2 GB of VRAM at 8 k context.
Does the model support image generation?
No, SmolVLM2 does not support image or video generation.
What is the context length?
The model supports up to 8192 tokens of context.