RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can I run Qwen3 Coder 30B A3B Instruct locally?

Qwen3 Coder 30B A3B Instruct by Qwen has 31B parameters (3.0B active per token, 128 experts, 8 used per token). At 4-bit with an 8K context it needs about 18.7 GB of GPU memory. It fits on 48 of 65 GPUs and Macs we track. See all models that fit in 24 GB of VRAM.

18.7 GBQ4_K_M · 8K context
RTX 3090smallest consumer GPU that fits
M4 32GBsmallest Mac that fits
256Kmax context
GQAattention type

Memory needed by quant and context length

GB of GPU memory = weights + KV cache (FP16) + runtime overhead. Weights use real file sizes from unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF where available (marked •), otherwise parameter count × bits per weight. How we calculate.

QuantQuality4K ctx8K ctx32K ctx128K ctx
2-bit (Q2_K) •rough11.6 GB11.9 GB14.2 GB23.2 GB
3-bit (Q3_K_M) •usable14.8 GB15.1 GB17.4 GB26.4 GB
4-bit (Q4_K_M) •good, most popular18.3 GB18.7 GB21.0 GB30.0 GB
5-bit (Q5_K_M) •very good21.3 GB21.7 GB23.9 GB32.9 GB
6-bit (Q6_K) •near lossless24.4 GB24.8 GB27.1 GB36.1 GB
8-bit (Q8_0) •effectively lossless31.3 GB31.7 GB33.9 GB42.9 GB
16-bit (F16/BF16) •original weights58.0 GB58.3 GB60.6 GB69.6 GB

Which GPUs and Macs run Qwen3 Coder 30B A3B Instruct?

Best quality quant that fits at 8K context. "Tight" = within 10% of usable memory. Speed is a ceiling from memory bandwidth (real speed is typically 50–80% of it).

HardwareMemoryFits?Best quantNeedsMax tok/s (est.)
RTX 3060 12GB12 GBNo———details
RTX 4060 Ti 16GB16 GBFitsQ2_K11.9 GB260details
RTX 309024 GBFitsQ4_K_M18.7 GB513details
RTX 409024 GBFitsQ4_K_M18.7 GB553details
RTX 509032 GBFitsQ6_K24.8 GB727details
RTX 5070 Ti16 GBFitsQ2_K11.9 GB810details
RX 7900 XTX24 GBFitsQ4_K_M18.7 GB527details
Mac M4 Pro 48GB48 GBFitsQ6_K24.8 GB111details
Mac M4 Max 128GB128 GBFitsF1658.3 GB91details
Mac M3 Ultra 512GB512 GBFitsF1658.3 GB136details
DGX Spark128 GBFitsF1658.3 GB45details
RTX 407012 GBNo———details
RTX 5060 Ti 16GB16 GBFitsQ2_K11.9 GB405details
RTX 40608 GBNo———details
Arc B58012 GBNo———details
RTX 3060 Ti8 GBNo———details
RTX 30708 GBNo———details
RTX 3080 10GB10 GBNo———details
RTX 3080 12GB12 GBNo———details
RTX 3090 Ti24 GBFitsQ4_K_M18.7 GB553details
RTX 4060 Ti 8GB8 GBNo———details
RTX 4070 Super12 GBNo———details
RTX 4070 Ti12 GBNo———details
RTX 4070 Ti Super16 GBFitsQ2_K11.9 GB607details
RTX 408016 GBFitsQ2_K11.9 GB648details
RTX 4080 Super16 GBFitsQ2_K11.9 GB665details
RTX 50608 GBNo———details
RTX 507012 GBNo———details
RTX 508016 GBFitsQ2_K11.9 GB868details
2x RTX 30902×24 GBFitsQ8_031.7 GB293details
2x RTX 40902×24 GBFitsQ8_031.7 GB316details
2x RTX 50902×32 GBFitsQ8_031.7 GB561details
RTX A600048 GBFitsQ8_031.7 GB241details
RTX 6000 Ada48 GBFitsQ8_031.7 GB301details
RTX PRO 6000 Blackwell96 GBFitsF1658.3 GB299details
L40S48 GBFitsQ8_031.7 GB271details
A100 40GB40 GBFitsQ8_031.7 GB487details
A100 80GB80 GBFitsF1658.3 GB340details
H100 PCIe80 GBFitsF1658.3 GB333details
H100 SXM80 GBFitsF1658.3 GB558details
H200141 GBFitsF1658.3 GB800details
RX 76008 GBNo———details
RX 7700 XT12 GBNo———details
RX 7800 XT16 GBFitsQ2_K11.9 GB564details
RX 7900 GRE16 GBFitsQ2_K11.9 GB521details
RX 7900 XT20 GBFitsQ3_K_M15.1 GB553details
RX 907016 GBFitsQ2_K11.9 GB579details
RX 9070 XT16 GBFitsQ2_K11.9 GB579details
Ryzen AI Max+ 395 128GB128 GBFitsF1658.3 GB43details
Arc A770 16GB16 GBFitsQ2_K11.9 GB506details
Mac M1 16GB16 GBNo———details
Mac M2 24GB24 GBFitsQ2_K11.9 GB90details
Mac M3 24GB24 GBFitsQ2_K11.9 GB90details
Mac M4 16GB16 GBNo———details
Mac M4 32GB32 GBFitsQ4_K_M18.7 GB66details
Mac M1 Pro 32GB32 GBFitsQ4_K_M18.7 GB110details
Mac M3 Pro 36GB36 GBFitsQ5_K_M21.7 GB70details
Mac M4 Pro 24GB24 GBFitsQ2_K11.9 GB247details
Mac M4 Pro 64GB64 GBFitsQ8_031.7 GB86details
Mac M1 Max 64GB64 GBFitsQ8_031.7 GB125details
Mac M2 Max 96GB96 GBFitsF1658.3 GB67details
Mac M3 Max 128GB128 GBFitsF1658.3 GB67details
Mac M4 Max 64GB64 GBFitsQ8_031.7 GB171details
Mac M2 Ultra 192GB192 GBFitsF1658.3 GB133details
Mac M3 Ultra 256GB256 GBFitsF1658.3 GB136details

Prefer an API?

Qwen: Qwen3 Coder 30B A3B Instruct is on OpenRouter at $0.15 per 1M input tokens and $0.60 per 1M output tokens. Price details and history →

Questions

How much VRAM does Qwen3 Coder 30B A3B Instruct need?

About 18.7 GB at 4-bit (Q4_K_M) with an 8K context: 17.3 GB of weights (real GGUF file size), 0.8 GB of KV cache and 0.7 GB of runtime overhead. At 8-bit it needs about 31.7 GB.

What is the smallest GPU that runs Qwen3 Coder 30B A3B Instruct?

At Q4_K_M and 8K context, the smallest single consumer GPU in our table that fits it is the RTX 3090 (24 GB).

Can I run Qwen3 Coder 30B A3B Instruct on a Mac?

Yes. At Q4_K_M with 8K context it fits on a Mac M4 32GB or larger, using the default macOS GPU memory limit.

Similar models

Same family or a similar parameter count, with the VRAM each needs at Q4_K_M and 8K context.

Qwen3 32B

33B · 21.1 GB at 4-bit

Qwen3.8 27B

28B · 16.8 GB at 4-bit

Qwen3.6 27B

28B · 17.0 GB at 4-bit

North Mini Code 1.0

30B · 19.3 GB at 4-bit

Hy Memory 30B A3B

30B · 18.5 GB at 4-bit

GLM 4.7 Flash

31B · 18.2 GB at 4-bit

Model facts

Hugging Face repoQwen/Qwen3-Coder-30B-A3B-Instruct
Parameters30,532,122,624 (31B) · ~3.0B active (from the model name)
Layers / hidden size48 / 2048
AttentionGQA; 32 heads, 4 KV heads, head dim 128
KV cache per token (FP16)96.0 KiB
Max context262,144 tokens
Licenseapache-2.0
Released on HFJul 31, 2025
Downloads (30d) / likes1,872,138 / 1,340

Sources: Hugging Face model API · config.json (Qwen/Qwen3-Coder-30B-A3B-Instruct) · GGUF files · GPU specs from vendor pages (list). Data fetched Nov 17, 2026. Spot an error? Report it.