Can I run Qwen3-Coder-30B-A3B-Instruct on an RTX 6000 Ada?

Yes, with a GGUF version

The published weights are too big, but the UD-Q8_K_XL file from unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF (36.0 GB) fits, with room for ≈ 72K tokens of context.

Full analysis of Qwen3-Coder-30B-A3B-Instruct Everything that runs on RTX 6000 Ada Get the UD-Q8_K_XL file on Hugging Face

ollama run hf.co/unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF:UD-Q8_K_XL

Files and context on RTX 6000 Ada

Each file of Qwen3-Coder-30B-A3B-Instruct (and of GGUF versions the site has analyzed), and how much context fits next to it, with the context cache in 16-bit or 8-bit.

File	Size	On RTX 6000 Ada	Context (16-bit)	Context (8-bit)
As published	61.1 GB	Doesn't fit	—	—
BF16unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	61.1 GB	Doesn't fit	—	—
UD-Q8_K_XLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	36.0 GB	Tight	72K	143K
Q8_0unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	32.5 GB	Fits	106K	213K
UD-Q6_K_XLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	26.3 GB	Fits	167K	256K
Q6_Kunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	25.1 GB	Fits	180K	256K
UD-Q5_K_XLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	21.7 GB	Fits	213K	256K
Q5_K_Munsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	21.7 GB	Fits	213K	256K
Q5_K_Sunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	21.1 GB	Fits	220K	256K
Q4_1unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	19.2 GB	Fits	238K	256K
Q4_K_Munsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	18.6 GB	Fits	245K	256K
UD-Q4_K_XLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	17.7 GB	Fits	254K	256K
Q4_K_Sunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	17.5 GB	Fits	256K	256K
Q4_0unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	17.4 GB	Fits	256K	256K
IQ4_NLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	17.3 GB	Fits	256K	256K
IQ4_XSunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	16.4 GB	Fits	256K	256K
Q3_K_Munsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	14.7 GB	Fits	256K	256K
UD-Q3_K_XLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	13.8 GB	Fits	256K	256K
Q3_K_Sunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	13.3 GB	Fits	256K	256K
UD-IQ3_XXSunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	12.8 GB	Fits	256K	256K
UD-Q2_K_XLunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	11.8 GB	Fits	256K	256K
Q2_K_Lunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	11.3 GB	Fits	256K	256K
Q2_Kunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	11.3 GB	Fits	256K	256K
UD-IQ2_Munsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	10.8 GB	Fits	256K	256K
UD-IQ2_XXSunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	10.3 GB	Fits	256K	256K
UD-IQ1_Munsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	9.63 GB	Fits	256K	256K
UD-IQ1_Sunsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	8.91 GB	Fits	256K	256K
UD-TQ1_0unsloth/Qwen3-Coder-30B-A3B-Instruct-GGUF	8.01 GB	Fits	256K	256K

Qwen3-Coder-30B-A3B-Instruct on other hardware

Popular models that run on RTX 6000 Ada

Assumes 90% of the card's memory is usable and at least 4K tokens of context for one conversation, with the context cache in 16-bit unless noted. Fits leaves at least 20% of that memory for context; Tight leaves less.