Qwen3.8 Whittle MoE 27B A17.8B needs about 20.0 GB of memory at Q4_K_M (a 17.4 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 7 quants, from 13.9 GB (Q3_K_M) to 28.7 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.
Whittle MoE 27B (A18B) – a 28B Qwen model with 17.8B active experts
Whittle MoE 27B (A18B) is a 28‑billion‑parameter Qwen model with 17.8 billion active experts. It supports vision and is released under Apache‑2.0. The model was quantized by logic65 and is available in several GGUF builds.
Good at
- Large context (262 144 tokens)
- Vision support
- Mixture‑of‑experts inference
Limits
- Requires high VRAM for full‑precision
- Large file sizes (up to 28.7 GB)
- No text‑only quantisation options beyond Q3_K_M
Best for
- Vision‑enabled LLM tasks
- Long‑form content generation
- Research on MoE architectures
Who it’s for. Developers and researchers needing a large‑scale, vision‑capable Qwen model with MoE. Suitable for environments with sufficient GPU memory and storage.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| Q3_K_M | 13.9 GB | 16.5 GB | Too big | Get file → | |
| — | 14.7 GB | 17.3 GB | Too big | Get file → | |
| Q4_K_M | 17.4 GB | 20.0 GB | Too big | Get file → | |
| — | 20.1 GB | 22.8 GB | Too big | Get file → | |
| — | 21.7 GB | 24.3 GB | Too big | Get file → | |
| Q6_K | 23.1 GB | 25.8 GB | Too big | Get file → | |
| Q8_0 | 28.7 GB | 31.3 GB | Too big | Get file → |
For 8 GB VRAM, use the Q3_K_M quant (13.87 GB file, ~16.5 GB memory). For 12 GB VRAM, Q4_K_M (17.37 GB, ~20 GB). For 16 GB VRAM, Q6_K (23.14 GB, ~25.8 GB). Higher quants give better quality but need more memory.
What GPU do you need for Qwen3.8 Whittle MoE 27B A17.8B?
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 the main file (20.1 GB file, about 22.8 GB in use).
- 32 GB VRAM: the best quant that fits is Q8_0 (28.7 GB file, about 31.3 GB in use).
Which Qwen3.8 Whittle MoE 27B A17.8B quant should you choose?
Q4_K_M is the usual starting point for Qwen3.8 Whittle MoE 27B A17.8B: it balances file size and quality well. Here it is a 17.4 GB file.
The Q8_0 file is 28.7 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 Q3_K_M at 13.9 GB, with the quality trade-offs that come with fewer bits. 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
Frequently asked questions about Qwen3.8 Whittle MoE 27B A17.8B
How much VRAM does Qwen3.8 Whittle MoE 27B A17.8B need?
About 20.0 GB at Q4_K_M with an 8K context window. That is the 17.4 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 Whittle MoE 27B A17.8B on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 16.5 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8 Whittle MoE 27B A17.8B GGUF?
It depends on the quant: Q3_K_M is 13.9 GB, file is 14.7 GB, Q4_K_M is 17.4 GB, file is 20.1 GB, file is 21.7 GB, Q6_K is 23.1 GB and 1 more.
What context length does Qwen3.8 Whittle MoE 27B A17.8B support?
The GGUF header lists a trained context length of 262,144 tokens (256K). The KV cache grows with the context you use: about 2.1 GB at 8K and 8.6 GB at 32K.
Where can I download Qwen3.8 Whittle MoE 27B A17.8B GGUF?
The files are on Hugging Face in the logic65/Qwen3.8-Whittle-MoE-27B-A17.8B-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 Whittle MoE 27B A17.8B 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 this model on 8 GB VRAM?
Use the Q3_K_M quant; it requires about 16.5 GB of VRAM at 8k context, so 8 GB is insufficient.
Does this model support vision?
Yes, the model has vision capability and includes an optional vision projector file.
What license applies?
The model is released under the Apache‑2.0 license.