Qwen3.8 27B Q3 K M needs about 16.2 GB of memory at Q3_K_M (a 13.5 GB file) with an 8K context window, so it fits on a 24 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.
Qwen3.8‑27B Q3_K_M – 27B vision‑language model in GGUF format
Qwen3.8‑27B Q3_K_M is a 27‑billion‑parameter vision‑language model quantized with Q3_K_M. The GGUF file is 13.5 GB and requires roughly 16.2 GB of VRAM at 8 k context. It supports image and video understanding and is compatible with llama.cpp.
Good at
- Vision‑language tasks
- Long‑context reasoning up to 262 144 tokens
- Agentic planning and tool use
Limits
- Requires ~18 GB VRAM for safe inference
- No fine‑tuning support in this release
- Apache‑2.0 license limits commercial use without attribution
Best for
- Image captioning and multimodal QA
- Complex multi‑step agentic workflows
- Research on vision‑language integration
Who it’s for. Developers needing a dense, deployment‑friendly vision‑language model for inference on local hardware. The model is suitable for research and production scenarios where a 27B parameter count is acceptable and the Apache‑2.0 license is compatible.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_M | 13.5 GB | 16.2 GB | Too big | Get file → |
For 8 GB VRAM, the Q3_K_M quant is too large; consider a higher‑compression quant if available. With 16 GB VRAM, Q3_K_M runs comfortably. For 32 GB VRAM, the model can be loaded with higher context or additional tools.
What GPU do you need for Qwen3.8 27B Q3 K M?
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: no quant fits at an 8K context.
- 6 GB VRAM: no quant fits at an 8K context.
- 8 GB VRAM: no quant fits at an 8K context.
- 12 GB VRAM: no quant fits at an 8K context.
- 16 GB VRAM: no quant fits at an 8K context.
- 24 GB VRAM: the best quant that fits is Q3_K_M (13.5 GB file, about 16.2 GB in use).
Which Qwen3.8 27B Q3 K M quant should you choose?
Q3_K_M (13.5 GB) is the middle option in this repository. A reasonable choice when a 4-bit file will not fit, mostly on larger models. Quality loss is noticeable but often acceptable.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3.8 27B Q3 K M
How much VRAM does Qwen3.8 27B Q3 K M need?
About 16.2 GB at Q3_K_M with an 8K context window. That is the 13.5 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 Qwen3.8 27B Q3 K M on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 16.2 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8 27B Q3 K M GGUF?
It depends on the quant: Q3_K_M is 13.5 GB.
What context length does Qwen3.8 27B Q3 K M support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.2 GB at 8K and 8.7 GB at 32K.
Where can I download Qwen3.8 27B Q3 K M GGUF?
The files are on Hugging Face in the Chungulus/Qwen3.8-27B-Q3_K_M-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 Qwen3.8 27B Q3 K M 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 Qwen3.8‑27B Q3_K_M on 8 GB VRAM?
No. The model requires approximately 18 GB of VRAM for safe inference.
Does this model support vision tasks?
Yes, it includes a native vision encoder for images and videos.
What license applies to this model?
The model is released under the Apache‑2.0 license.