QW

Qwen3.8 27B Abliterated GSQ Orca

QwenGGUF by ayozheeUpdated 2026-09-27
Download on Hugging Face
Parameters27.3B
Context256K
Layers65
Architectureqwen35
Smallest quant9.6 GB
Quick answer

Qwen3.8 27B Abliterated GSQ Orca needs about 12.3 GB of memory at IQ2_S (a 9.6 GB file) with an 8K context window, so it fits on a 16 GB GPU. This repository has 2 quants, from 9.6 GB (IQ2_S) to 12.1 GB (IQ3_S). Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.

Qwen3.8-27B Abliterated GSQ-Orca GGUF with non‑uniform quantization

Qwen3.8-27B Abliterated GSQ-Orca is a 27 billion‑parameter model quantized with GSQ and RCO. It offers a vision projector and optional MTP head for speculative decoding. The quantizations are non‑uniform, allocating precision per tensor under a size budget.

Good at

  • Non‑uniform low‑bit quantization
  • Vision projector support
  • MTP head for speculative decoding

Limits

  • Requires 12–14 GB VRAM at 8 k context
  • No other vision or audio modalities
  • Apache‑2.0 license only

Best for

  • Local inference on mid‑range GPUs
  • Multimodal text‑image tasks
  • Speculative decoding in llama.cpp

Who it’s for. Developers and researchers needing a compact Qwen3.8‑27B model for local inference. The model is suitable for users with RTX 3060‑style GPUs and those who want to experiment with selective tensor replacement and abliterated weights.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
IQ2_S9.6 GB 12.3 GB
Too big Get file →
IQ3_S12.1 GB 14.8 GB
Too big Get file →

For 12 GB VRAM, use IQ2_S (9.61 GB, ~12.3 GB VRAM at 8 k context). If more memory is available, IQ3_S (12.12 GB, ~14.8 GB VRAM) gives higher precision. Both are approximate estimates.

What GPU do you need for Qwen3.8 27B Abliterated GSQ Orca?

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 IQ3_S (12.1 GB file, about 14.8 GB in use).

Which Qwen3.8 27B Abliterated GSQ Orca quant should you choose?

IQ2_S (9.6 GB) is the middle option in this repository. 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

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

Source

Quantized byayozhee
FamilyQwen
Licenseapache-2.0
Downloads (HF)9,096
Last updated2026-09-27

Frequently asked questions about Qwen3.8 27B Abliterated GSQ Orca

How much VRAM does Qwen3.8 27B Abliterated GSQ Orca need?

About 12.3 GB at IQ2_S with an 8K context window. That is the 9.6 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 Abliterated GSQ Orca on an 8 GB GPU?

Not at an 8K context. The smallest file needs about 12.3 GB, so you would need a larger GPU or CPU offloading.

What is the file size of Qwen3.8 27B Abliterated GSQ Orca GGUF?

It depends on the quant: IQ2_S is 9.6 GB, IQ3_S is 12.1 GB.

What context length does Qwen3.8 27B Abliterated GSQ Orca 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 Abliterated GSQ Orca GGUF?

The files are on Hugging Face in the ayozhee/Qwen3.8-27B-Abliterated-GSQ-Orca-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 Abliterated GSQ Orca 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?

No, the smallest quantization requires about 12 GB VRAM at 8 k context.

Does the model support vision tasks?

Yes, it includes a vision projector for multimodal use.

Is the MTP head necessary for generation?

No, the MTP head is optional and only needed for speculative decoding.

Other GGUFs of Qwen3.8-27B-GSQ-RCO-GGUF

Similar models

Browse all →

More Qwen models

All Qwen →

Other medium models

All medium models →