Skip to content
localmodel.run

Text model · Qwen3.6

Qwen3.6 35B-A3B: RAM and VRAM requirements

Qwen3.6 35B-A3B needs about 24.5 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~22.29 GB to download; KV cache and overhead add the rest), or about 39.1 GB at Q8_0. The lightest hardware that runs it is Nvidia GeForce RTX 5090 (32GB).

Qwen3.6 family · 36B params (Mixture-of-Experts: activates only 3B of 36B params per token, so generation is faster than the total size suggests) · released Apr 2026.

Shopping for hardware? See what runs Qwen3.6 35B-A3B →

License Apache-2.0 · Commercial OK ↓ 5.9M/mo ♥ 2.6K on HuggingFace
Q4_K_M GGUF
22.29 GB
Q8_0 GGUF
36.9 GB
Memory @ Q4 (4k)
~24.5 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?

Qwen3.6 35B-A3B runs on 9 of 40 tracked devices at Q4_K_M.

9 run well 0 tight fit 31 too small
Fit check Q4_K_M
No, not enough memory
needs 24.5 GB usable 10.5 GB
See the full breakdown

Memory breakdown

Weights (Q4_K_M)22.29 GB
+
KV cache (4k)1.4 GB
+
Overhead0.8 GB
=
Total24.5 GB

How context length changes it

4k context ~24.5 GB 32k context ~34.6 GB 128k context ~69.2 GB

Longer context grows the KV cache, which for Qwen3.6 35B-A3B is sized by its 3B active params, not the full 36B. It needs ~24.5 GB at 4k and ~69.2 GB at 128k.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 15.1 GB est
Q3_K_M 17.6 GB est
Q4_K_M (default) 22.29 GB
Q5_K_M 25.7 GB est
Q6_K 29.5 GB est
Q8_0 36.9 GB
FP16 69.38 GB

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

Run it

Ollama
$ ollama run qwen3.6:35b-a3b
llama.cpp
$ llama-cli -hf bartowski/Qwen_Qwen3.6-35B-A3B-GGUF:Q4_K_M
LM Studio
$ lms get bartowski/Qwen_Qwen3.6-35B-A3B-GGUF

Which devices can run Qwen3.6 35B-A3B?

Apple Silicon Macs

RAM-only laptops

iPhone & iPad

Android

NVIDIA GPUs

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 Ornith 1.0 35B 35B Next size up Mixtral 8x7B 46.7B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does Qwen3.6 35B-A3B need?

At Q4_K_M, Qwen3.6 35B-A3B needs about 24.5 GB (weights ~22.29 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~39.1 GB.

What is the Q4_K_M GGUF file size of Qwen3.6 35B-A3B?

The Q4_K_M GGUF file is about 22.29 GB to download, and the Q8_0 GGUF is about 36.9 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~24.5 GB of memory at Q4_K_M.

Can Qwen3.6 35B-A3B run on a laptop?

Qwen3.6 35B-A3B is large; you need a high-memory Mac or multi-GPU setup at Q4_K_M.

Is Qwen3.6 35B-A3B cheaper to run because it is a MoE model?

It is faster, not lighter. Qwen3.6 35B-A3B activates only 3B of 36B params per token (so it runs quickly), but all experts must stay in memory, so it still needs memory for the full 36B.

Can I use Qwen3.6 35B-A3B commercially?

Yes. Qwen3.6 35B-A3B is licensed Apache-2.0, which permits commercial use.

Understand the numbers

Short guides to the ideas behind Qwen3.6 35B-A3B's memory and quant figures.

Successor to Qwen3 30B-A3B: 35B total, 3B active (256 experts, 8 active per token). Multimodal, 256K context. Q4_K_M from bartowski, Q8_0/BF16 from unsloth; Ollama default pulls 24GB. Fast on unified-memory machines thanks to the 3B active params.

Sources

Last validated 2026-08-03. Memory figures are estimates. See methodology.