Home / Will it run? / Qwen3 30B-A3B
Alibaba Qwen · Qwen3
Will Qwen3 30B-A3B run on my GPU?
30.5B parameters (mixture of experts, 3.3B active per token, but every expert must sit in memory), 32,768-token context, Apache-2.0. At Q4_K_M the weights come to about 17.0 GiB; add 1.5 GiB for the runtime and you need roughly 18.5 GiB of VRAM before context.
Weights by quant
GiBQ4_K_M · the everyday quant17.0 GiB
Q8_0 · near-lossless30.1 GiB
FP16 · full weights56.8 GiB
Runtime reserve1.5 GiB
Card by card
4 of 18 fit at Q4Headroom is what remains for the KV cache after weights and the 1.5 GiB reserve. How far it stretches depends on the architecture: grouped-query models are frugal, older dense models are not. If the table says tight, plan on a shorter context or a lower quant. Parameter count from the model's own card. The Buy column opens live Amazon listings; those links may earn us a commission.