Models › Qwen3-30B-A3B

Qwen3-30B-A3B: VRAM needs and speed by GPU

At Q4_K_M, Qwen3-30B-A3B needs about 20.1 GB of VRAM. No measured runs yet. Speeds below are estimates. Submit a run.

30.53B parameters (mixture-of-experts, ~3.1B active) · license: apache-2.0 · 40,960-token context · GGUF published Apr 28, 2025 · GGUF files on Hugging Face

Memory needed per quantization

QuantFile sizeNeeds (estimate)Smallest card it fits
UD-IQ1_S9.0 GB10.6 GB12 GB cards
UD-IQ1_M9.7 GB11.2 GB12 GB cards
UD-IQ2_XXS10.4 GB11.9 GB16 GB cards
UD-IQ2_M10.9 GB12.4 GB16 GB cards
Q2_K11.3 GB12.8 GB16 GB cards
Q2_K_L11.3 GB12.9 GB16 GB cards
UD-Q2_K_XL11.8 GB13.4 GB16 GB cards
UD-IQ3_XXS12.9 GB14.4 GB16 GB cards
Q3_K_S13.3 GB14.9 GB16 GB cards
UD-Q3_K_XL13.8 GB15.4 GB16 GB cards
Q3_K_M14.7 GB16.3 GB24 GB cards
IQ4_XS16.4 GB17.9 GB24 GB cards
IQ4_NL17.3 GB18.9 GB24 GB cards
Q4_017.4 GB18.9 GB24 GB cards
Q4_K_S17.5 GB19.0 GB24 GB cards
UD-Q4_K_XL17.7 GB19.3 GB24 GB cards
Q4_K_M18.6 GB20.1 GB24 GB cards
Q4_119.2 GB20.8 GB24 GB cards
Q5_K_S21.1 GB22.6 GB24 GB cards
Q5_K_M21.7 GB23.3 GB24 GB cards
UD-Q5_K_XL21.7 GB23.3 GB24 GB cards
Q6_K25.1 GB26.7 GB32 GB cards
UD-Q6_K_XL26.3 GB27.9 GB32 GB cards
Q8_032.5 GB34.0 GB48 GB cards
UD-Q8_K_XL36.0 GB37.5 GB48 GB cards
BF1661.1 GB62.7 GB96 GB cards

Expected speed on popular hardware

HardwareBest quant that fitsExpected tok/s
NVIDIA RTX 3060UD-IQ1_M159 estimate
NVIDIA RTX 4060 TiDoes not fit
NVIDIA RTX 3090Q5_K_M151 estimate
NVIDIA RTX 4090Q5_K_M167 estimate
NVIDIA RTX 5090Q6_K229 estimate
AMD Radeon RX 7900 XTXQ5_K_M168 estimate
Apple M4 Pro (20-core GPU) (64 GB)Q8_035.6 estimate
Apple M3 Max (40-core GPU) (64 GB)Q8_052.0 estimate
AMD Ryzen AI Max+ 395 (Radeon 8060S)Q8_032.0 estimate

Check your own card

Too big for your card?

Run Qwen3-30B-A3B on a rented 24-80 GB GPU by the hour.

Rent a bigger GPU on RunPod

Affiliate link: if you sign up, RigBench may earn a referral credit. It never affects rankings or numbers.