Can I run Ornith-1.5-35B-A3B on an NVIDIA H100?

Yes, with a GGUF version

The published weights are too big, but the BF16 file from ornith-ai/Ornith-1.5-35B-A3B-GGUF (71.1 GB) fits, with room for ≈ 41K tokens of context.

Full analysis of Ornith-1.5-35B-A3B Everything that runs on NVIDIA H100 Get the BF16 file on Hugging Face

ollama run hf.co/ornith-ai/Ornith-1.5-35B-A3B-GGUF:BF16

Files and context on NVIDIA H100

Each file of Ornith-1.5-35B-A3B (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)
As published	71.9 GB	Doesn't fit	2K	3K
BF16ornith-ai/Ornith-1.5-35B-A3B-GGUF	71.1 GB	Tight	41K	83K
Q8_0ornith-ai/Ornith-1.5-35B-A3B-GGUF	37.8 GB	Fits	256K	256K
Q6_Kornith-ai/Ornith-1.5-35B-A3B-GGUF	29.2 GB	Fits	256K	256K
Q5_K_Mornith-ai/Ornith-1.5-35B-A3B-GGUF	25.3 GB	Fits	256K	256K
Q4_K_Mornith-ai/Ornith-1.5-35B-A3B-GGUF	21.7 GB	Fits	256K	256K

Ornith-1.5-35B-A3B 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.