QW

Qwen3.8 27B Flash MTP

QwenGGUF by SuperLogicUpdated 2026-10-24
Download on Hugging Face
Parameters461M
Context256K
Layers65
Architectureqwen35
Smallest quant6.2 GB
Quick answer

Qwen3.8 27B Flash MTP needs about 8.9 GB of memory at its main file (a 6.2 GB file) with an 8K context window, so it fits on a 12 GB GPU. Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.

Qwen3.8 27B Flash MTP – a 27B ternary code model with MTP self‑speculation for 8 GB VRAM

Qwen3.8 27B Flash MTP is a 27B parameter model quantised to ternary. It includes a self‑speculative draft head and native vision‑language support. The model runs on consumer GPUs with 8 GB VRAM and is optimized for code and editing tasks.

Good at

  • Code generation
  • Vision‑language understanding
  • Interactive drafting

Limits

  • Requires 8 GB VRAM or more
  • No external vision API needed
  • License is apache‑2.0

Best for

  • Developers needing fast code completion on laptops
  • Researchers working with vision‑language datasets
  • Teams building local inference services

Who it’s for. This model is suited for users who want a large, dense LLM that fits on a single 8 GB GPU. It is ideal for developers, researchers, and small teams that require local inference without cloud costs.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
—6.2 GB 8.9 GB
Too big Get file →

The provided GGUF file is ternary and occupies 6.21 GB. It requires about 8.9 GB of VRAM at 8 k context. For 8 GB VRAM, use this quant. If more VRAM is available, the same file can be used; no other quantisations are listed.

What GPU do you need for Qwen3.8 27B Flash 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: the best quant that fits is the main file (6.2 GB file, about 8.9 GB in use).

Specifications

Architecture from the GGUF header

Architectureqwen35
Parameters461M
Layers65
Trained context262,144 tokens
KV cache (fp16)260 KB per token

Source

Quantized bySuperLogic
FamilyQwen
Licenseapache-2.0
Downloads (HF)5,010
Last updated2026-10-24

Frequently asked questions about Qwen3.8 27B Flash MTP

How much VRAM does Qwen3.8 27B Flash MTP need?

About 8.9 GB at its main file with an 8K context window. That is the 6.2 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 Flash MTP on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 8.9 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Qwen3.8 27B Flash MTP GGUF?

It depends on the quant: file is 6.2 GB.

What context length does Qwen3.8 27B Flash 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 2.2 GB at 8K and 8.7 GB at 32K.

Where can I download Qwen3.8 27B Flash MTP GGUF?

The files are on Hugging Face in the SuperLogic/Qwen3.8-27B-Flash-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 Flash 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.

Does the model support vision tasks?

Yes, it includes native vision‑language understanding for images and videos.

What license applies to this model?

The model is released under the apache‑2.0 license.

Other GGUFs of Qwen3.8-27B

Similar models

Browse all →

More from SuperLogic

All by SuperLogic →

More Qwen models

All Qwen →

Other tiny models

All tiny models →