LLMap
Models /Gemma 3n 4B /Can I run it?
Compatibility check

Can I run Gemma 3n 4B on L40S?

Yes — it runs.best quant:
Decode speed
146 tok/sest
Usable context
33K
Memory
48 GB
Fits in memory146 tok/sest · 33K ctx
Fits in memory125 tok/sest · 33K ctx
Fits in memory83 tok/sest · 33K ctx
What is quantisation? →
Gemma 3n 4B
7.8B · full specs & benchmarks →
L40S
48 GB · everything it runs →