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

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

Yes

Its Q8_0 file (29.0 GB) fits in the 72.0 GB an NVIDIA H100 can use, with room for ≈ 256K tokens of context.

Full analysis of Qwen3.8-27B-Uncensored-GGUF Everything that runs on NVIDIA H100 Get the Q8_0 file on Hugging Face
ollama run hf.co/JonathanColetti/Qwen3.8-27B-Uncensored-GGUF:Q8_0

Files and context on NVIDIA H100

Each file of Qwen3.8-27B-Uncensored-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 H100 Context (16-bit)Context (8-bit)
Q8_029.0 GBFits256K256K
Qwen3.8-27B-Uncensored-noMTP-Q8_028.6 GBFits256K256K
Q6_K22.4 GBFits256K256K
Qwen3.8-27B-Uncensored-noMTP-Q6_K22.1 GBFits256K256K
Q5_K_M19.5 GBFits256K256K
Qwen3.8-27B-Uncensored-noMTP-Q5_K_M19.2 GBFits256K256K
Q4_K_M16.8 GBFits256K256K
Qwen3.8-27B-Uncensored-noMTP-Q4_K_M16.5 GBFits256K256K
IQ4_XS15.3 GBFits256K256K
Qwen3.8-27B-Uncensored-noMTP-IQ4_XS15.1 GBFits256K256K
IQ2_M10.6 GBFits256K256K
Qwen3.8-27B-Uncensored-noMTP-IQ2_M10.2 GBFits256K256K

Qwen3.8-27B-Uncensored-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.