Can I run Dirk-Qwen3.8-27B-GGUF on an NVIDIA H100?

Yes

Its UD-Q8_K_XL file (31.5 GB) fits in the 72.0 GB an NVIDIA H100 can use, with room for ≈ 256K tokens of context.

Full analysis of Dirk-Qwen3.8-27B-GGUF Everything that runs on NVIDIA H100 Get the UD-Q8_K_XL file on Hugging Face

ollama run hf.co/peculiar-ragdoll/Dirk-Qwen3.8-27B-GGUF:UD-Q8_K_XL

Files and context on NVIDIA H100

Each file of Dirk-Qwen3.8-27B-GGUF (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 NVIDIA H100	Context (16-bit)	Context (8-bit)
UD-Q8_K_XL	31.5 GB	Fits	256K	256K
UD-Q8_K_L	28.0 GB	Fits	256K	256K
UD-Q6_K_XL	25.3 GB	Fits	256K	256K
UD-Q6_K	22.0 GB	Fits	256K	256K
UD-Q5_K_XL	20.9 GB	Fits	256K	256K
UD-Q4_K_XL	17.6 GB	Fits	256K	256K
UD-Q4_K_S	15.4 GB	Fits	256K	256K
UD-IQ4_XS	14.3 GB	Fits	256K	256K
UD-Q3_K_XL	13.1 GB	Fits	256K	256K
IQ3_S	12.1 GB	Fits	256K	256K
IQ3_XXS	10.4 GB	Fits	256K	256K
UD-Q2_K_XL	9.83 GB	Fits	256K	256K
IQ2_S	9.61 GB	Fits	256K	256K
IQ2_XS	8.77 GB	Fits	256K	256K

Dirk-Qwen3.8-27B-GGUF on other hardware

Popular models that run on NVIDIA H100

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.