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

Can I run Qwen3-VL-32B-Thinking-GGUF on an NVIDIA H200?

Yes

Its F16 file (65.5 GB) fits in the 127 GB an NVIDIA H200 can use, with room for ≈ 229K tokens of context.

Full analysis of Qwen3-VL-32B-Thinking-GGUF Everything that runs on NVIDIA H200 Get the F16 file on Hugging Face
llama-server -hf Qwen/Qwen3-VL-32B-Thinking-GGUF:F16

Files and context on NVIDIA H200

Each file of Qwen3-VL-32B-Thinking-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)
F1665.5 GBFits229K256K
Q8_034.8 GBFits256K256K
Q4_K_M19.8 GBFits256K256K

Qwen3-VL-32B-Thinking-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.