Skip to content

Can L40S 48G run Nemotron 3.5 Lightning 30B-A3B?

Comfortable: GGUF Q8_0 needs 35.1 GB at 4K, batch 1; headroom 12.9 GB.

Text input, FP16 KV cache, one card. Unified-memory capacity uses the GPU guide’s usable limit; negative headroom may require host offload. Speed source settings may differ from this 4K estimate.

Measured runs

No measured run for this pair. Sizing is an estimate, not a runtime test.

Quant memory at 4K

QuantTotal GBVerdictWeight sizing
GGUF Q4_019.8ComfortableVerified file
GGUF Q8_035.1ComfortableVerified file

Context capacity · GGUF Q8_0

ContextTotal GBVerdict
4K35.1Comfortable
16K35.2Comfortable
32K35.3Comfortable
128K35.9Comfortable

Generation speed

GGUF Q8_0: 285 tok/s · Bandwidth ceiling (not measured)

Other model / GPU pairs