QW

Merkyor/Qwen3.8-27B-Coder390-EfficientThink

QwenGGUF by gsdavidpUpdated 2026-10-14
Download on Hugging Face
Parameters27.3B
Context256K
Layers65
Architectureqwen35
Smallest quant3.2 GB
Quick answer

Merkyor/Qwen3.8-27B-Coder390-EfficientThink needs about 20.2 GB of memory at its main file (a 17.5 GB file) with an 8K context window, so it fits on a 24 GB GPU. This repository has 9 quants, from 3.2 GB (Q8_0) to 29.1 GB (Q8_0). Longer context needs more: at 32K tokens the KV cache alone adds about 8.7 GB.

Qwen3.8-27B-Coder390-EfficientThink – 27.3B multimodal LLM

This model is a 27.3B parameter Qwen architecture trained for coding, reasoning, and vision tasks. It supports a 262,144 token context window and is released under the Apache‑2.0 license. The model is available in multiple GGUF quantizations.

Good at

  • Coding assistance
  • Multimodal reasoning
  • Large context handling

Limits

  • No strict loops observed in evaluations
  • Requires large VRAM for full precision
  • Vision support only in certain quantizations

Best for

  • Developers needing code generation
  • Researchers working with long documents
  • Applications requiring multimodal input

Who it’s for. Engineers and data scientists who need a powerful, open‑source LLM for code, text, and image tasks. The model is suitable for projects that can accommodate its memory requirements and support the Apache‑2.0 license.

Quants & memory

QuantFileEst. memoryvs VRAMResultDownload
Q8_03.2 GB 5.8 GB
Fits Get file →
—13.3 GB 16.0 GB
Too big Get file →
—13.5 GB 16.2 GB
Too big Get file →
—17.3 GB 20.0 GB
Too big Get file →
—17.5 GB 20.2 GB
Too big Get file →
—19.6 GB 22.3 GB
Too big Get file →
—19.8 GB 22.5 GB
Too big Get file →
Q6_K23.2 GB 25.9 GB
Too big Get file →
Q8_029.1 GB 31.8 GB
Too big Get file →

For 8 GB VRAM, use the Q6_K quantization (23.18 GB file, ~25.9 GB memory at 8k context). For 16 GB VRAM, the Q8_0 quantization (3.16 GB file, ~5.8 GB memory) is recommended. Larger VRAM allows use of the full‑precision Q8_0 (29.07 GB file, ~31.8 GB memory).

What GPU do you need for Merkyor/Qwen3.8-27B-Coder390-EfficientThink?

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: the best quant that fits is Q8_0 (3.2 GB file, about 5.8 GB in use).
  • 8 GB VRAM: the best quant that fits is Q8_0 (3.2 GB file, about 5.8 GB in use).
  • 12 GB VRAM: the best quant that fits is Q8_0 (3.2 GB file, about 5.8 GB in use).
  • 16 GB VRAM: the best quant that fits is the main file (13.3 GB file, about 16.0 GB in use).
  • 24 GB VRAM: the best quant that fits is the main file (19.8 GB file, about 22.5 GB in use).
  • 32 GB VRAM: the best quant that fits is Q8_0 (29.1 GB file, about 31.8 GB in use).

Which Merkyor/Qwen3.8-27B-Coder390-EfficientThink quant should you choose?

If memory is tight, the smallest file is Q8_0 at 3.2 GB, with the quality trade-offs that come with fewer bits. Close to lossless. It costs roughly twice the memory of a 4-bit file, so it suits cases where you have plenty of VRAM.

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

Frequently asked questions about Merkyor/Qwen3.8-27B-Coder390-EfficientThink

How much VRAM does Merkyor/Qwen3.8-27B-Coder390-EfficientThink need?

About 20.2 GB at its main file with an 8K context window. That is the 17.5 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 Merkyor/Qwen3.8-27B-Coder390-EfficientThink on an 8 GB GPU?

Yes, with Q8_0: the 3.2 GB file needs about 5.8 GB at 8K context. Larger quants in this repository do not fit in 8 GB.

What is the file size of Merkyor/Qwen3.8-27B-Coder390-EfficientThink GGUF?

It depends on the quant: Q8_0 is 3.2 GB, file is 13.3 GB, file is 13.5 GB, file is 17.3 GB, file is 17.5 GB, file is 19.6 GB and 3 more.

What context length does Merkyor/Qwen3.8-27B-Coder390-EfficientThink 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 Merkyor/Qwen3.8-27B-Coder390-EfficientThink GGUF?

The files are on Hugging Face in the gsdavidp/Qwen3.8-27B-Coder390 repository. The download buttons on this page open the file on Hugging Face. ggufmodels.com does not host any model files.

What license does Merkyor/Qwen3.8-27B-Coder390-EfficientThink 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 the Q8_0 quantization on 8 GB VRAM?

The Q8_0 quantization requires approximately 5.8 GB memory at 8k context, which fits within 8 GB VRAM with some headroom.

What is the memory requirement for the Q6_K quantization?

The Q6_K quantization estimates about 25.9 GB memory at 8k context, so it needs a GPU with at least 26 GB VRAM.

Does the model support vision tasks?

Yes, the model has detected vision abilities, but vision support may depend on the specific quantization and deployment environment.

Other GGUFs of Qwen3.8-27B-EfficientThink-Uncensored-K3-Opus5-Grok4.6-GPT5.6Sol-SFT-SimPO-DFlash2

Similar models

Browse all →

More Qwen models

All Qwen →

Other medium models

All medium models →