Can I run Qwen3 30B-A3B locally?
Yes. At Q4 with 8K context, Qwen3 30B-A3B needs about 19 GB, so it runs entirely on any GPU with 20 GB of VRAM or more, starting with the AMD Radeon RX 7900 XT.
Memory needed by compression and context length
| Compression | 4K context | 8K context | 32K context |
|---|---|---|---|
Q4 The most common download. Small quality loss. | 19 GB | 19 GB | 21 GB |
Q5 Very close to full quality. | 22 GB | 22 GB | 24 GB |
Q8 Practically indistinguishable from FP16 at about half the size. | 32 GB | 32 GB | 34 GB |
FP16 Uncompressed. Reference quality, largest size. | 58 GB | 59 GB | 61 GB |
Total memory: model weights, context (KV cache) and about 1 GB of runtime overhead. To run entirely on the GPU, VRAM must be about 5% larger than this.
GPUs that run Qwen3 30B-A3B entirely on the GPU
- Amazon
AMD Radeon RX 7900 XT
20 GB · ~197 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 3090 Ti
24 GB · ~231 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 3090
24 GB · ~220 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 4090
24 GB · ~231 tok/s at Q4
- Amazon
AMD Radeon RX 7900 XTX
24 GB · ~223 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 4090 D
24 GB · ~231 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 5090 D
32 GB · ~323 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 5090
32 GB · ~323 tok/s at Q4
- Amazon
NVIDIA GeForce RTX 5090 D v2
32 GB · — at Q4
Model facts
- Parameters
- 30.5B (3.3B active, mixture-of-experts)
- Max context
- 32K tokens
- License
- Apache 2.0
- Released
- Apr 2025
Check it with your hardware
Change the compression, context length and system RAM in the calculator.
Open calculator