Text model · Qwen3
Qwen3 30B-A3B: RAM and VRAM requirements
Qwen3 30B-A3B needs about 20.7 GB to run at Q4_K_M (the Q4_K_M GGUF file is ~18.6 GB to download; KV cache and overhead add the rest), or about 34.6 GB at Q8_0. The lightest hardware that runs it is Nvidia GeForce RTX 4090 (24GB).
Qwen3 family · 30.5B params (Mixture-of-Experts: activates only 3.3B of 30.5B params per token, so generation is faster than the total size suggests) · released Apr 2025 · 23M Ollama pulls · LMArena Elo 1383.
Shopping for hardware? See what runs Qwen3 30B-A3B →
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 30B-A3B runs on 13 of 40 tracked devices at Q4_K_M.
Memory breakdown
How context length changes it
Longer context grows the KV cache, which for Qwen3 30B-A3B is sized by its 3.3B active params, not the full 30.5B. It needs ~20.7 GB at 4k and ~61.8 GB at 128k.
Benchmark scores
Sourced third-party benchmarks for the full-precision Qwen3 30B-A3B, not a specific quant. See the full leaderboard.
Quantization sizes
| Quantization | Size on disk |
|---|---|
| Q2_K | 12.8 GB est |
| Q3_K_M | 14.9 GB est |
| Q4_K_M (default) | 18.6 GB |
| Q5_K_M | 21.7 GB est |
| Q6_K | 25 GB est |
| Q8_0 | 32.5 GB |
| FP16 | 61 GB |
Lower quant = smaller and faster, slightly lower quality. Q4_K_M is the common default.
Run it
ollama run qwen3:30b-a3b llama-cli -hf unsloth/Qwen3-30B-A3B-GGUF:Q4_K_M lms get unsloth/Qwen3-30B-A3B-GGUF Which devices can run Qwen3 30B-A3B?
Apple Silicon Macs
- Apple M1 (8GB) No
- Apple M2 (16GB) No
- Apple M4 (16GB) No
- Apple M5 (16GB) No
- Apple M3 Pro (18GB) No
- Apple M4 (24GB) No
- Apple M4 Pro (24GB) No
- Apple M5 (32GB) Tight
- Apple M4 Pro (48GB) Yes
- Apple M5 Pro (48GB) Yes
- Apple M4 Max (64GB) Yes
- Apple M4 Max (128GB) Yes
- Apple M5 Max (128GB) Yes
- Apple M3 Ultra (256GB) Yes
RAM-only laptops
iPhone & iPad
Android
NVIDIA GPUs
Same job, different size
Pick by what fits your memory: step down to free up VRAM, or step up if you have headroom.
Similar models
Head-to-head
FAQ
How much VRAM or RAM does Qwen3 30B-A3B need?
At Q4_K_M, Qwen3 30B-A3B needs about 20.7 GB (weights ~18.6 GB + KV cache + overhead) at a 4k context. At Q8_0 budget ~34.6 GB.
What is the Q4_K_M GGUF file size of Qwen3 30B-A3B?
The Q4_K_M GGUF file is about 18.6 GB to download, and the Q8_0 GGUF is about 32.5 GB. That is the weights file on disk; to run it you also need room for the KV cache and overhead, so budget ~20.7 GB of memory at Q4_K_M.
Can Qwen3 30B-A3B run on a laptop?
Qwen3 30B-A3B is large; you need a 24 GB+ GPU or a 32-48 GB Mac at Q4_K_M.
Is Qwen3 30B-A3B cheaper to run because it is a MoE model?
It is faster, not lighter. Qwen3 30B-A3B activates only 3.3B of 30.5B params per token (so it runs quickly), but all experts must stay in memory, so it still needs memory for the full 30.5B.
Can I use Qwen3 30B-A3B commercially?
Yes. Qwen3 30B-A3B is licensed Apache-2.0, which permits commercial use.
Understand the numbers
Short guides to the ideas behind Qwen3 30B-A3B's memory and quant figures.
MoE: 30.5B total / 3.3B active (128 total experts, 8 activated per token). Q4_K_M=18.6GB and Q8_0=32.5GB from official unsloth/Qwen3-30B-A3B-GGUF HF repo, cross-confirmed with Qwen/Qwen3-30B-A3B-GGUF. Native context 32K, extendable to 128K via YaRN. Despite large Q4 size, inference is fast due to only 3.3B active params.
Sources
Last validated 2026-08-03. Memory figures are estimates. See methodology.