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

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

Yes, with a GGUF version

The published weights are too big, but the Q8_0 file from Qwen/Qwen3-235B-A22B-GGUF (250 GB) fits, with room for ≈ 40K tokens of context.

Full analysis of Qwen3-235B-A22B Everything that runs on NVIDIA B300 Get the Q8_0 file on Hugging Face
llama-server -hf Qwen/Qwen3-235B-A22B-GGUF:Q8_0

Files and context on NVIDIA B300

Each file of Qwen3-235B-A22B (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 B300 Context (16-bit)Context (8-bit)
As published470 GBDoesn't fit——
Q8_0Qwen/Qwen3-235B-A22B-GGUF250 GBTight40K40K
Q6_KQwen/Qwen3-235B-A22B-GGUF193 GBFits40K40K
Q5_K_MQwen/Qwen3-235B-A22B-GGUF167 GBFits40K40K
Q5_0Qwen/Qwen3-235B-A22B-GGUF162 GBFits40K40K
Q4_K_MQwen/Qwen3-235B-A22B-GGUF142 GBFits40K40K

Qwen3-235B-A22B 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.