Can I run Qwen3-VL-235B-A22B-Thinking on an NVIDIA B300?

Yes, with a GGUF version

The published weights are too big, but the Q8_0 file from Qwen/Qwen3-VL-235B-A22B-Thinking-GGUF (250 GB) fits, with room for ≈ 47K tokens of context.

Full analysis of Qwen3-VL-235B-A22B-Thinking Everything that runs on NVIDIA B300 Get the Q8_0 file on Hugging Face

llama-server -hf Qwen/Qwen3-VL-235B-A22B-Thinking-GGUF:Q8_0

Files and context on NVIDIA B300

Each file of Qwen3-VL-235B-A22B-Thinking (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 B300	Context (16-bit)	Context (8-bit)
As published	471 GB	Doesn't fit	—	—
F16Qwen/Qwen3-VL-235B-A22B-Thinking-GGUF	470 GB	Doesn't fit	—	—
Q8_0Qwen/Qwen3-VL-235B-A22B-Thinking-GGUF	250 GB	Tight	47K	94K
Q4_K_MQwen/Qwen3-VL-235B-A22B-Thinking-GGUF	142 GB	Fits	256K	256K

Qwen3-VL-235B-A22B-Thinking on other hardware

Popular models that run on NVIDIA B300

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.