Can I run Qwen2.5-32B-Instruct on an RTX 6000 Ada?

Yes, with a GGUF version

The published weights are too big, but the Q8_0 file from bartowski/Qwen2.5-32B-Instruct-GGUF (34.8 GB) fits, with room for ≈ 31K tokens of context.

Full analysis of Qwen2.5-32B-Instruct Everything that runs on RTX 6000 Ada Get the Q8_0 file on Hugging Face

ollama run hf.co/bartowski/Qwen2.5-32B-Instruct-GGUF:Q8_0

Files and context on RTX 6000 Ada

Each file of Qwen2.5-32B-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	65.5 GB	Doesn't fit	—	—
F16bartowski/Qwen2.5-32B-Instruct-GGUF	65.5 GB	Doesn't fit	—	—
Q8_0bartowski/Qwen2.5-32B-Instruct-GGUF	34.8 GB	Tight	31K	32K
Q6_K_Lbartowski/Qwen2.5-32B-Instruct-GGUF	27.3 GB	Fits	32K	32K
Q6_Kbartowski/Qwen2.5-32B-Instruct-GGUF	26.9 GB	Fits	32K	32K
Q5_K_Lbartowski/Qwen2.5-32B-Instruct-GGUF	23.7 GB	Fits	32K	32K
Q5_K_Mbartowski/Qwen2.5-32B-Instruct-GGUF	23.3 GB	Fits	32K	32K
Q5_K_Sbartowski/Qwen2.5-32B-Instruct-GGUF	22.6 GB	Fits	32K	32K
Q4_K_Lbartowski/Qwen2.5-32B-Instruct-GGUF	20.4 GB	Fits	32K	32K
Q4_K_Mbartowski/Qwen2.5-32B-Instruct-GGUF	19.9 GB	Fits	32K	32K
Q4_K_Sbartowski/Qwen2.5-32B-Instruct-GGUF	18.8 GB	Fits	32K	32K
Q4_0bartowski/Qwen2.5-32B-Instruct-GGUF	18.7 GB	Fits	32K	32K
Qwen2.5-32B-Instruct-Q4_0_4_4bartowski/Qwen2.5-32B-Instruct-GGUF	18.6 GB	Fits	32K	32K
Qwen2.5-32B-Instruct-Q4_0_4_8bartowski/Qwen2.5-32B-Instruct-GGUF	18.6 GB	Fits	32K	32K
Qwen2.5-32B-Instruct-Q4_0_8_8bartowski/Qwen2.5-32B-Instruct-GGUF	18.6 GB	Fits	32K	32K
Q3_K_XLbartowski/Qwen2.5-32B-Instruct-GGUF	17.9 GB	Fits	32K	32K
IQ4_XSbartowski/Qwen2.5-32B-Instruct-GGUF	17.7 GB	Fits	32K	32K
Q3_K_Lbartowski/Qwen2.5-32B-Instruct-GGUF	17.2 GB	Fits	32K	32K
Q3_K_Mbartowski/Qwen2.5-32B-Instruct-GGUF	15.9 GB	Fits	32K	32K
IQ3_Mbartowski/Qwen2.5-32B-Instruct-GGUF	14.8 GB	Fits	32K	32K
Q3_K_Sbartowski/Qwen2.5-32B-Instruct-GGUF	14.4 GB	Fits	32K	32K
IQ3_XSbartowski/Qwen2.5-32B-Instruct-GGUF	13.7 GB	Fits	32K	32K
Q2_K_Lbartowski/Qwen2.5-32B-Instruct-GGUF	13.1 GB	Fits	32K	32K
Q2_Kbartowski/Qwen2.5-32B-Instruct-GGUF	12.3 GB	Fits	32K	32K
IQ2_Mbartowski/Qwen2.5-32B-Instruct-GGUF	11.3 GB	Fits	32K	32K
IQ2_Sbartowski/Qwen2.5-32B-Instruct-GGUF	10.4 GB	Fits	32K	32K
IQ2_XSbartowski/Qwen2.5-32B-Instruct-GGUF	9.96 GB	Fits	32K	32K
IQ2_XXSbartowski/Qwen2.5-32B-Instruct-GGUF	9.03 GB	Fits	32K	32K

Qwen2.5-32B-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.