Compatibility check
Can I run Qwen2.5 VL 72B Instruct on H100 80GB SXM?
Yes — it runs.best quant:
Decode speed
60 tok/sest
Usable context
66K
Memory
80 GB
Fits in memory60 tok/sest · 66K ctx
Fits in memory51 tok/sest · 66K ctx
Spills to system RAMest