Model Anatomy See what's inside open-weight AI models

Can I run Kolibri-1-GGUF on an NVIDIA H200?

Yes

Its Q8_0 file (83.1 GB) fits in the 127 GB an NVIDIA H200 can use, with room for ≈ 256K tokens of context.

Full analysis of Kolibri-1-GGUF Everything that runs on NVIDIA H200 Get the Q8_0 file on Hugging Face
llama-server -hf Hob-forge/Kolibri-1-GGUF:Q8_0

Files and context on NVIDIA H200

Each file of Kolibri-1-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.

FileSizeOn NVIDIA H200 Context (16-bit)Context (8-bit)
Q8_083.1 GBFits256K256K
Q6_K64.2 GBFits256K256K
Q5_K_M55.6 GBFits256K256K
Q4_K_M47.5 GBFits256K256K
Q3_K_M37.5 GBFits256K256K
Q2_K28.6 GBFits256K256K

Kolibri-1-GGUF on other hardware

Popular models that run on NVIDIA H200

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.