Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP needs about 20.7 GB of memory at Q4_K_M (a 18.0 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 20 quants, from 11.7 GB (IQ2_M) to 30.2 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.6 GB.
Qwen3.8 27B Cold Fusion GAIN V1.1 – a 27‑billion‑parameter model with reduced reasoning tokens
This 27B‑parameter Qwen model uses the Cold Fusion (GAIN+Unsloth) training method to keep 99 % of BF16 performance while cutting thinking tokens to 1/10–1/2 of the original. It supports coding, reasoning, roleplay, and vision tasks. The model is released under Apache‑2.0.
Good at
- Efficient reasoning with fewer tokens
- Fast multi‑token processing (MTP)
- Versatile coding and vision capabilities
Limits
- Large context size requires significant VRAM
- No explicit vision support in GGUF format
- Requires 8‑bit or 4‑bit quantization for practical use
Best for
- Developers needing fast, detailed code generation
- Researchers testing reasoning efficiency
- Teams deploying LLMs on GPUs with limited VRAM
Who it’s for. Engineers and researchers who need a high‑detail, low‑token‑usage model for coding, reasoning, and roleplay tasks. The model is suitable for GPU‑based inference with 8‑bit or 4‑bit quantization.
Quants & memory
| Quant | File | Est. memory | vs VRAM | Result | |
|---|---|---|---|---|---|
| IQ2_M | 11.7 GB | 14.3 GB | Too big | Get file → | |
| IQ2_M | 12.1 GB | 14.8 GB | Too big | Get file → | |
| IQ3_M | 14.1 GB | 16.7 GB | Too big | Get file → | |
| IQ3_M | 14.5 GB | 17.2 GB | Too big | Get file → | |
| IQ4_XS | 15.3 GB | 18.0 GB | Too big | Get file → | |
| IQ4_XS | 16.6 GB | 19.2 GB | Too big | Get file → | |
| IQ4_XS | 17.0 GB | 19.7 GB | Too big | Get file → | |
| Q4_K_S | 17.1 GB | 19.7 GB | Too big | Get file → | |
| Q4_K_S | 17.5 GB | 20.2 GB | Too big | Get file → | |
| Q4_K_M | 18.0 GB | 20.7 GB | Too big | Get file → | |
| Q4_K_M | 18.5 GB | 21.1 GB | Too big | Get file → | |
| Q5_K_S | 20.2 GB | 22.8 GB | Too big | Get file → | |
| Q5_K_S | 20.6 GB | 23.3 GB | Too big | Get file → | |
| Q5_K_M | 20.7 GB | 23.4 GB | Too big | Get file → | |
| Q5_K_M | 21.2 GB | 23.8 GB | Too big | Get file → | |
| Q6_K | 22.4 GB | 25.1 GB | Too big | Get file → | |
| Q6_K | 23.6 GB | 26.2 GB | Too big | Get file → | |
| Q6_K | 24.0 GB | 26.7 GB | Too big | Get file → | |
| Q8_0 | 29.8 GB | 32.4 GB | Too big | Get file → | |
| Q8_0 | 30.2 GB | 32.9 GB | Too big | Get file → |
For 8 GB VRAM, use Q4_K_M (≈20.7 GB memory at 8 k context). For 16 GB VRAM, Q5_K_S (≈22.8 GB) or Q6_K (≈25.1 GB) provide a balance of quality and memory. Larger GPUs can run Q8_0 for maximum fidelity.
What GPU do you need for Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP?
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: the best quant that fits is IQ2_M (12.1 GB file, about 14.8 GB in use).
- 24 GB VRAM: the best quant that fits is Q5_K_M (21.2 GB file, about 23.8 GB in use).
- 32 GB VRAM: the best quant that fits is Q6_K (24.0 GB file, about 26.7 GB in use).
- 48 GB VRAM: the best quant that fits is Q8_0 (30.2 GB file, about 32.9 GB in use).
Which Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP quant should you choose?
Q4_K_M is the usual starting point for Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP: it balances file size and quality well. Here it is a 18.0 GB file.
The Q8_0 file is 29.8 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 IQ2_M at 11.7 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
Frequently asked questions about Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP
How much VRAM does Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP need?
About 20.7 GB at Q4_K_M with an 8K context window. That is the 18.0 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 Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP on an 8 GB GPU?
Not at an 8K context. The smallest file needs about 14.3 GB, so you would need a larger GPU or CPU offloading.
What is the file size of Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP GGUF?
It depends on the quant: IQ2_M is 11.7 GB, IQ2_M is 12.1 GB, IQ3_M is 14.1 GB, IQ3_M is 14.5 GB, IQ4_XS is 15.3 GB, IQ4_XS is 16.6 GB and 14 more.
What context length does Qwen3.8 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP 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 27B Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP GGUF?
The files are on Hugging Face in the DavidAU/Qwen3.8-27B-Cold-Fusion-GAIN-V1.1-NM-DAU-NEO-MAX-MTP-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 Cold Fusion GAIN V1.1 NM DAU NEO MAX MTP 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 Q4_K_M quant, which estimates 20.7 GB memory at 8 k context – it will not fit on 8 GB VRAM, so a higher‑quant or smaller context is needed.
Does the model support vision tasks?
The model detects vision ability, but the GGUF files do not include vision weights; vision support is not available in this format.
What license applies to this model?
The model is released under the Apache‑2.0 license, allowing free use and modification.