QW

Swift-1.5-Qwen3.8-27B-Uncensored-MTP

QwenGGUF by bjivanovichUpdated 2026-10-04
Download on Hugging Face
Parameters27.3B
Context256K
Layers65
Architectureqwen35
Smallest quant17.3 GB
Quick answer

Swift-1.5-Qwen3.8-27B-Uncensored-MTP needs about 20.0 GB of memory at Q4_K_M (a 17.3 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 4 quants, from 17.3 GB (Q4_K_M) to 29.0 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.

Swift-1.5-Qwen3.8-27B-Uncensored-MTP – a 27.3B Qwen model with embedded MTP and vision support

Swift-1.5-Qwen3.8-27B-Uncensored-MTP is a 27.3B parameter Qwen model fine‑tuned for reasoning efficiency and refusal ablation. It retains the vision tower and includes an embedded Multi‑Token Prediction head for speculative decoding. The model is released under Apache‑2.0.

Good at

  • Vision‑enabled reasoning
  • Embedded MTP draft decoding
  • Uncensored content generation

Limits

  • Requires high VRAM for full context
  • No support for non‑linear quants in llamAmpere
  • Large context length may exceed typical VRAM

Best for

  • Vision‑based question answering
  • High‑throughput inference on RTX 3090/4090
  • Draft‑based speculative decoding

Who it’s for. Developers building vision‑enabled LLM applications who need fast, speculative decoding and a large context window. The model is suitable for research and production environments that can accommodate its memory requirements.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q4_K_M17.3 GB 20.0 GB
Too big Get file →
Q5_K_M20.2 GB 22.8 GB
Too big Get file →
Q6_K22.4 GB 25.1 GB
Too big Get file →
Q8_029.0 GB 31.7 GB
Too big Get file →

For 8 GB VRAM, use Q4_K_M (file size 17.33 GB, estimated 20 GB at 8 k context). For 16 GB VRAM, Q5_K_M (20.15 GB, 22.8 GB). For 24 GB VRAM, Q6_K (22.43 GB, 25.1 GB) or Q8_0 (29.05 GB, 31.7 GB). Memory estimates are approximate.

What GPU do you need for Swift-1.5-Qwen3.8-27B-Uncensored-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: no quant fits at an 8K context.
  • 24 GB VRAM: the best quant that fits is Q5_K_M (20.2 GB file, about 22.8 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (29.0 GB file, about 31.7 GB in use).

Which Swift-1.5-Qwen3.8-27B-Uncensored-MTP quant should you choose?

Q4_K_M is the usual starting point for Swift-1.5-Qwen3.8-27B-Uncensored-MTP: it balances file size and quality well. Here it is a 17.3 GB file.

The Q8_0 file is 29.0 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.

New to quant names? Read GGUF quantization explained.

Specifications

Architecture from the GGUF header

Architectureqwen35
Parameters27.3B
Layers65
Trained context262,144 tokens
KV cache (fp16)260 KB per token

Source

Quantized bybjivanovich
FamilyQwen
Licenseapache-2.0
Downloads (HF)850
Last updated2026-10-04

Frequently asked questions about Swift-1.5-Qwen3.8-27B-Uncensored-MTP

How much VRAM does Swift-1.5-Qwen3.8-27B-Uncensored-MTP need?

About 20.0 GB at Q4_K_M with an 8K context window. That is the 17.3 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 Swift-1.5-Qwen3.8-27B-Uncensored-MTP on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 20.0 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Swift-1.5-Qwen3.8-27B-Uncensored-MTP GGUF?

It depends on the quant: Q4_K_M is 17.3 GB, Q5_K_M is 20.2 GB, Q6_K is 22.4 GB, Q8_0 is 29.0 GB.

What context length does Swift-1.5-Qwen3.8-27B-Uncensored-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 Swift-1.5-Qwen3.8-27B-Uncensored-MTP GGUF?

The files are on Hugging Face in the bjivanovich/ATX-Swift-1.5-Qwen3.8-27B-Uncensored-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 Swift-1.5-Qwen3.8-27B-Uncensored-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 Swift-1.5-Qwen3.8-27B-Uncensored-MTP on 8 GB VRAM?

The Q4_K_M quant requires about 20 GB at 8 k context, so 8 GB VRAM is insufficient.

Does the model support vision tasks?

Yes, the model includes a vision tower and is detected as vision‑capable.

What license applies to this model?

The model is released under the Apache‑2.0 license.

Other GGUFs of Swift-1.5-Qwen3.8-27B-Uncensored-MTP

Similar models

Browse all →

More from bjivanovich

All by bjivanovich →

More Qwen models

All Qwen →

Other medium models

All medium models →