Text model · Qwen3

Qwen3 32B: RAM and VRAM requirements

Qwen3 32B needs about 22 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~19.8 GB to download; KV cache and overhead add the rest), or about 37 GB at Q8_0. The lightest hardware that runs it is Nvidia GeForce RTX 4090 (24GB).

Qwen3 family · 32B params · released Apr 2025 · 23M Ollama pulls · LMArena Elo 1347.

Shopping for hardware? See what runs Qwen3 32B →

License Apache-2.0 · Commercial OK ↓ 8.5M/mo ♥ 727 on HuggingFace

Q4_K_M GGUF

19.8 GB

Q8_0 GGUF

34.8 GB

Memory @ Q4 (4k)

~22 GB

Context

32 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 32B runs on 12 of 40 tracked devices at Q4_K_M.

9 run well 3 tight fit 28 too small

Fit check Q4_K_M

Your device

Apple M4 (16GB)

No, not enough memory

needs 22 GB usable 10.5 GB

See the full breakdown

Memory breakdown

Weights (Q4_K_M)19.8 GB

KV cache (4k)1.4 GB

Overhead0.8 GB

Total22 GB

How context length changes it

4k context ~22 GB 32k context ~31.5 GB 128k context ~64 GB

Longer context grows the KV cache quickly: Qwen3 32B needs ~22 GB at 4k but ~64 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 32B generates at roughly ~31% of its short-context speed. How this is estimated.

Benchmark scores

Aider polyglot · coding

#7 of 10 ranked · Percent of 225 Exercism exercises solved across six languages.

40%

BFCL · tool use

#2 of 19 ranked · Overall Acc on the Berkeley Function-Calling Leaderboard.

48.7%

LMArena · chat (elo)

#4 of 16 ranked · Human-preference Elo from LMArena's blind head-to-head votes.

1347

Sourced third-party benchmarks for the full-precision Qwen3 32B, not a specific quant. See the full leaderboard.

Quantization sizes

GGUF quantson disk

Quantization	Size on disk
Q2_K	13.4 GB est
Q3_K_M	15.6 GB est
Q4_K_M (default)	19.8 GB
Q5_K_M	22.8 GB est
Q6_K	26.2 GB est
Q8_0	34.8 GB
FP16	66 GB

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

Run it

Ollama

$ ollama run qwen3:32b

llama.cpp

$ llama-cli -hf Qwen/Qwen3-32B-GGUF:Q4_K_M

LM Studio

$ lms get Qwen/Qwen3-32B-GGUF

Which devices can run Qwen3 32B?

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 North Mini Code 1.0 30.5B Next size up Gemma 4 31B 32.7B

Similar models

Head-to-head

FAQ

How much VRAM or RAM does Qwen3 32B need?

At Q4_K_M, Qwen3 32B needs about 22 GB (weights ~19.8 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~37 GB.

What is the Q4_K_M GGUF file size of Qwen3 32B?

The Q4_K_M GGUF file is about 19.8 GB to download, and the Q8_0 GGUF is about 34.8 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~22 GB of memory at Q4_K_M.

Can Qwen3 32B run on a laptop?

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

Can I use Qwen3 32B commercially?

Yes. Qwen3 32B is licensed Apache-2.0, which permits commercial use.

Understand the numbers

Short guides to the ideas behind Qwen3 32B's memory and quant figures.

Official Qwen/Qwen3-32B-GGUF HF repo lists Q4_K_M=19.8GB, Q8_0=34.8GB. Cross-checked with Ollama tags page (20GB rounded). Context 32K native, 128K via YaRN.

Sources

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