Can I run Ornith-1.5-122B on an NVIDIA B200?

Yes, with a GGUF version

The published weights are too big, but the Q8_0 file from unsloth/Ornith-1.5-122B-GGUF (133 GB) fits, with room for ≈ 256K tokens of context.

Full analysis of Ornith-1.5-122B Everything that runs on NVIDIA B200 Get the Q8_0 file on Hugging Face

llama-server -hf unsloth/Ornith-1.5-122B-GGUF:Q8_0

Files and context on NVIDIA B200

Each file of Ornith-1.5-122B (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 B200	Context (16-bit)	Context (8-bit)
As published	250 GB	Doesn't fit	—	—
BF16unsloth/Ornith-1.5-122B-GGUF	249 GB	Doesn't fit	—	—
Q8_0unsloth/Ornith-1.5-122B-GGUF	133 GB	Tight	256K	256K
UD-Q6_K_XLunsloth/Ornith-1.5-122B-GGUF	114 GB	Fits	256K	256K
UD-Q5_K_XLunsloth/Ornith-1.5-122B-GGUF	94.6 GB	Fits	256K	256K
UD-Q4_K_XLunsloth/Ornith-1.5-122B-GGUF	79.7 GB	Fits	256K	256K
UD-Q3_K_XLunsloth/Ornith-1.5-122B-GGUF	59.8 GB	Fits	256K	256K
UD-IQ3_XXSunsloth/Ornith-1.5-122B-GGUF	49.5 GB	Fits	256K	256K
UD-Q2_K_XLunsloth/Ornith-1.5-122B-GGUF	44.5 GB	Fits	256K	256K
UD-IQ2_XXSunsloth/Ornith-1.5-122B-GGUF	42.0 GB	Fits	256K	256K
UD-IQ1_Munsloth/Ornith-1.5-122B-GGUF	40.4 GB	Fits	256K	256K
BF16ornith-ai/Ornith-1.5-122B-GGUF	249 GB	Doesn't fit	—	—
Q8_0ornith-ai/Ornith-1.5-122B-GGUF	133 GB	Tight	256K	256K
Q6_Kornith-ai/Ornith-1.5-122B-GGUF	102 GB	Fits	256K	256K
Q5_K_Mornith-ai/Ornith-1.5-122B-GGUF	88.7 GB	Fits	256K	256K
Q4_K_Mornith-ai/Ornith-1.5-122B-GGUF	75.8 GB	Fits	256K	256K

Ornith-1.5-122B on other hardware

Popular models that run on NVIDIA B200

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.