Qwen
Qwen3.5-397B-A17B
image text to textapache-2.0836.7K downloadsHugging Face
No server in stock right now has the 902 GB this model needs for inference.
GPU memory it needs (estimate)
Rule of thumb from the model's size. Real needs depend on batch size, context length and framework.
Inference (full precision)
902 GB
Inference (8-bit)
451 GB
Inference (4-bit)
225 GB
LoRA fine-tuning
977 GB
QLoRA (8-bit base)
488 GB
Full fine-tuning
3607 GB