Skip to content
localmodel.run

Text model · Qwen3.8

Qwen3.8 27B: RAM and VRAM requirements

Qwen3.8 27B needs about 17.3 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~15.33 GB to download; KV cache and overhead add the rest), or about 29.1 GB at Q8_0. The lightest hardware that runs it is Nvidia GeForce RTX 4090 (24GB).

Qwen3.8 family · 27B params · released Aug 2026 · 1.1M Ollama pulls.

Shopping for hardware? See what runs Qwen3.8 27B →

License Apache-2.0 · Commercial OK ↓ 6.4M/mo ♥ 14.2K on HuggingFace
Q4_K_M GGUF
15.33 GB
Q8_0 GGUF
27.05 GB
Memory @ Q4 (4k)
~17.3 GB
Context
256 k

Q4_K_M and Q8_0 are GGUF download (file) sizes. Memory @ Q4 includes the KV cache and overhead, so it runs larger than the file on disk.

Will it run on your device?

Qwen3.8 27B runs on 13 of 43 tracked devices at Q4_K_M.

13 run well 0 tight fit 30 too small
Fit check Q4_K_M
No, not enough memory
needs 17.3 GB usable 10.5 GB
See the full breakdown

Memory breakdown

Weights (Q4_K_M)15.33 GB
+
KV cache (4k)1.2 GB
+
Overhead0.8 GB
=
Total17.3 GB

How context length changes it

4k context ~17.3 GB 32k context ~26.1 GB 128k context ~56 GB

Longer context grows the KV cache quickly: Qwen3.8 27B needs ~17.3 GB at 4k but ~56 GB at 128k, which can push it past a device that fits it at a short context.

Speed drops too: every token re-reads the KV cache, so at 128k context Qwen3.8 27B generates at roughly ~28% of its short-context speed. How this is estimated.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 11.3 GB est
Q3_K_M 13.2 GB est
Q4_K_M (default) 15.33 GB
Q5_K_M 19.2 GB est
Q6_K 22.1 GB est
Q8_0 27.05 GB
FP16 55.59 GB

Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.

Run it

Ollama
$ ollama run qwen3.8:27b
llama.cpp
$ llama-cli -hf unsloth/Qwen3.8-27B-GGUF:Q4_K_M
LM Studio
$ lms get unsloth/Qwen3.8-27B-GGUF

Which devices can run Qwen3.8 27B?

NVIDIA GPUs

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

AMD GPUs

Same job, different size

Pick by what fits your memory: step down to free up VRAM, or step up if you have headroom.

Next size down Gemma 4 26B-A4B 26.5B Next size up Qwen3.6 27B 27.8B

Similar models

FAQ

How much VRAM or RAM does Qwen3.8 27B need?

At Q4_K_M, Qwen3.8 27B needs about 17.3 GB (weights ~15.33 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~29.1 GB.

What is the Q4_K_M GGUF file size of Qwen3.8 27B?

The Q4_K_M GGUF file is about 15.33 GB to download, and the Q8_0 GGUF is about 27.05 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~17.3 GB of memory at Q4_K_M.

Can Qwen3.8 27B run on a laptop?

Qwen3.8 27B is large; you need a 24 GB+ GPU or a 32-48 GB Mac at Q4_K_M.

Can I use Qwen3.8 27B commercially?

Yes. Qwen3.8 27B is licensed Apache-2.0, which permits commercial use.

Understand the numbers

Short guides to the ideas behind Qwen3.8 27B's memory and quant figures.

Dense 27B with native vision (Qwen3.5-generation VLM architecture), 256K context, Apache-2.0. Q4_K_M, Q8_0 and BF16 sums from unsloth's repo; Ollama's default tag bundles the vision projector (17.74GB). MTP variant tags exist for speculative-decoding speedups.

Sources

Last validated 2026-09-07. Memory figures are estimates. See methodology.