Skip to content
localmodel.run

Text model · Qwen3-Coder

Qwen3-Coder 480B-A35B Instruct: RAM and VRAM requirements

Qwen3-Coder 480B-A35B Instruct needs about 276.2 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~270.14 GB to download; KV cache and overhead add the rest), or about 481.4 GB at Q8_0. The lightest hardware that runs it is a high-memory machine.

Qwen3-Coder family · 480B params (Mixture-of-Experts: activates only 35B of 480B params per token, so generation is faster than the total size suggests) · released Jul 2025.

Shopping for hardware? See what runs Qwen3-Coder 480B-A35B Instruct →

License apache-2.0 · Commercial OK ↓ 91.3K/mo ♥ 1.4K on HuggingFace
Q4_K_M GGUF
270.14 GB
Q8_0 GGUF
475.3 GB
Memory @ Q4 (4k)
~276.2 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-Coder 480B-A35B Instruct runs on 0 of 40 tracked devices at Q4_K_M.

0 run well 0 tight fit 40 too small
Fit check Q4_K_M
No, not enough memory
needs 276.2 GB usable 10.5 GB
See the full breakdown

Memory breakdown

Weights (Q4_K_M)270.14 GB
+
KV cache (4k)5.3 GB
+
Overhead0.8 GB
=
Total276.2 GB

How context length changes it

4k context ~276.2 GB 32k context ~313 GB 128k context ~439.2 GB

Longer context grows the KV cache, which for Qwen3-Coder 480B-A35B Instruct is sized by its 35B active params, not the full 480B. It needs ~276.2 GB at 4k and ~439.2 GB at 128k.

Quantization sizes

GGUF quantson disk
Quantization Size on disk
Q2_K 201 GB est
Q3_K_M 234.6 GB est
Q4_K_M (default) 270.14 GB
Q5_K_M 342 GB est
Q6_K 393.6 GB est
Q8_0 475.3 GB
FP16 960 GB

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

Run it

Ollama
$ ollama run qwen3-coder:480b
llama.cpp
$ llama-cli -hf unsloth/Qwen3-Coder-480B-A35B-Instruct-GGUF:Q4_K_M
LM Studio
$ lms get unsloth/Qwen3-Coder-480B-A35B-Instruct-GGUF

Which devices can run Qwen3-Coder 480B-A35B Instruct?

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 MiniMax-M1-80k 456B Next size up Nemotron 3 Ultra 550B-A55B 550B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does Qwen3-Coder 480B-A35B Instruct need?

At Q4_K_M, Qwen3-Coder 480B-A35B Instruct needs about 276.2 GB (weights ~270.14 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~481.4 GB.

What is the Q4_K_M GGUF file size of Qwen3-Coder 480B-A35B Instruct?

The Q4_K_M GGUF file is about 270.14 GB to download, and the Q8_0 GGUF is about 475.3 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~276.2 GB of memory at Q4_K_M.

Can Qwen3-Coder 480B-A35B Instruct run on a laptop?

Qwen3-Coder 480B-A35B Instruct is large; you need a high-memory Mac or multi-GPU setup at Q4_K_M.

Is Qwen3-Coder 480B-A35B Instruct cheaper to run because it is a MoE model?

It is faster, not lighter. Qwen3-Coder 480B-A35B Instruct activates only 35B of 480B params per token (so it runs quickly), but all experts must stay in memory, so it still needs memory for the full 480B.

Can I use Qwen3-Coder 480B-A35B Instruct commercially?

Yes. Qwen3-Coder 480B-A35B Instruct is licensed apache-2.0, which permits commercial use.

Understand the numbers

Short guides to the ideas behind Qwen3-Coder 480B-A35B Instruct's memory and quant figures.

Qwen3's flagship coding MoE: 480B total / 35B active, 256K context. Server-class hardware; the 30B-A3B sibling runs on consumer machines.

Sources

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