Qwen3.8-27B needs about 19.2 GB of memory at Q4_K_M (a 16.6 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 38 quants, from 6.6 GB to 54.7 GB (BF16). Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.
Qwen3.8-27B – 27.3B parameter vision‑language model
Qwen3.8-27B is a 27.3B parameter dense model with native vision‑language support. It is built on the Qwen3.5 architecture and offers improved coding, professional work, research, and long‑horizon agentic tasks. The model is released under the apache‑2.0 license.
Good at
- Vision‑language understanding
- Long‑horizon agentic tasks
- Coding and research support
Limits
- No vision support in older quantizations
- Large VRAM requirement for high‑quality quants
- Requires llama.cpp fork for ternary hybrids
Best for
- Complex multi‑step task completion
- Image and video analysis
- Agent planning and execution
Who it’s for. Developers and researchers needing a dense vision‑language model for complex reasoning and agentic workflows. The model is suitable for local deployment with sufficient VRAM and the appropriate llama.cpp fork for ternary quant files.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| — | 6.6 GB | 9.3 GB | Too big | Get file → | |
| — | 7.7 GB | 10.4 GB | Too big | Get file → | |
| IQ2_S | 8.4 GB | 11.1 GB | Too big | Get file → | |
| IQ2_S | 8.8 GB | 11.5 GB | Too big | Get file → | |
| Q2_K_XL | 10.0 GB | 12.7 GB | Too big | Get file → | |
| IQ3_XXS | 10.1 GB | 12.8 GB | Too big | Get file → | |
| IQ3_XXS | 10.4 GB | 13.1 GB | Too big | Get file → | |
| Q2_K | 10.9 GB | 13.5 GB | Too big | Get file → | |
| IQ3_XXS | 11.0 GB | 13.7 GB | Too big | Get file → | |
| IQ3_XXS | 11.4 GB | 14.0 GB | Too big | Get file → | |
| IQ3_S | 11.8 GB | 14.5 GB | Too big | Get file → | |
| IQ3_S | 12.0 GB | 14.6 GB | Too big | Get file → | |
| IQ3_S | 12.1 GB | 14.8 GB | Too big | Get file → | |
| IQ3_S | 12.3 GB | 15.0 GB | Too big | Get file → | |
| Q3_K_XL | 13.3 GB | 16.0 GB | Too big | Get file → | |
| Q3_K | 13.5 GB | 16.2 GB | Too big | Get file → | |
| IQ4_XS | 14.4 GB | 17.1 GB | Too big | Get file → | |
| Q4_K_S | 15.6 GB | 18.3 GB | Too big | Get file → | |
| Q4_K_M | 16.6 GB | 19.2 GB | Too big | Get file → | |
| Q4_K | 16.8 GB | 19.5 GB | Too big | Get file → | |
| Q2_K_L | 17.2 GB | 19.9 GB | Too big | Get file → | |
| Q4_K_XL | 17.4 GB | 20.1 GB | Too big | Get file → | |
| Q3_K_L | 18.7 GB | 21.4 GB | Too big | Get file → | |
| Q5_K | 19.5 GB | 22.2 GB | Too big | Get file → | |
| Q5_K_XL | 20.7 GB | 23.4 GB | Too big | Get file → | |
| Q4_K_L | 20.9 GB | 23.6 GB | Too big | Get file → | |
| Q6_K | 22.4 GB | 25.1 GB | Too big | Get file → | |
| Q5_K_L | 22.9 GB | 25.6 GB | Too big | Get file → | |
| Q6_K_L | 23.9 GB | 26.6 GB | Too big | Get file → | |
| Q6_K_XL | 24.8 GB | 27.5 GB | Too big | Get file → | |
| Q6_K_L | 24.9 GB | 27.6 GB | Too big | Get file → | |
| Q8_K_L | 27.3 GB | 29.9 GB | Too big | Get file → | |
| Q8_0 | 29.0 GB | 31.7 GB | Too big | Get file → | |
| Q8_K_XL | 31.5 GB | 34.1 GB | Too big | Get file → | |
| Q8_0 | 38.8 GB | 41.5 GB | Too big | Get file → | |
| F16 | 53.8 GB | 56.5 GB | Too big | Get file → | |
| BF16 | 54.7 GB | 57.3 GB | Too big | Get file → | |
| BF16 | 54.7 GB | 57.3 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_S (15.62 GB file, 18.3 GB memory at 8 k context). For 16 GB VRAM, Q4_K_M (16.55 GB file, 19.2 GB memory) or Q4_K (16.81 GB file, 19.5 GB memory). For 32 GB VRAM, Q8_K_XL (31.46 GB file, 34.1 GB memory) or Q8_0 (38.78 GB file, 41.5 GB memory).
What GPU do you need for Qwen3.8-27B?
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: the best quant that fits is IQ2_S (8.8 GB file, about 11.5 GB in use).
- 16 GB VRAM: the best quant that fits is Q3_K_XL (13.3 GB file, about 16.0 GB in use).
- 24 GB VRAM: the best quant that fits is Q4_K_L (20.9 GB file, about 23.6 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (29.0 GB file, about 31.7 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (38.8 GB file, about 41.5 GB in use).
Which Qwen3.8-27B quant should you choose?
Q4_K_M is the usual starting point for Qwen3.8-27B: it balances file size and quality well. Here it is a 16.6 GB file.
The Q8_0 file is 29.0 GB, about 1.8 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 the first one in the table at 6.6 GB, with the quality trade-offs that come with fewer bits.
New to quant names? Read GGUF quantization explained.
Specifications
Architecture from the GGUF header
Source
Frequently asked questions about Qwen3.8-27B
How much VRAM does Qwen3.8-27B need?
About 19.2 GB at Q4_K_M with an 8K context window. That is the 16.6 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 on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 9.3 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8-27B GGUF?
It depends on the quant: file is 6.6 GB, file is 7.7 GB, IQ2_S is 8.4 GB, IQ2_S is 8.8 GB, Q2_K_XL is 10.0 GB, IQ3_XXS is 10.1 GB and 32 more.
What context length does Qwen3.8-27B 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 GGUF?
The files are on Hugging Face in the Sh200126/Huihui-Qwen3.8-27B-abliterated-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 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 on 8 GB VRAM?
Use the Q4_K_S quantization; it requires about 18.3 GB memory at 8 k context, which is close to 8 GB VRAM but may need off‑loading.
Does the model support vision tasks?
Yes, it is a native vision‑language model that understands images and videos.
Do I need a special llama.cpp fork?
Only for ternary hybrid‑attention kernels; standard llama.cpp will not run those files.