Skip to content

Can H100 80G run Nemotron 3.5 Bolt 74B-A9B?

Comfortable: GGUF IQ2_XXS needs 23.1 GB at 4K, batch 1; headroom 57.0 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 IQ2_XXS23.1ComfortableVerified file

Context capacity · GGUF IQ2_XXS

ContextTotal GBVerdict
4K23.1Comfortable
16K23.1Comfortable
32K23.3Comfortable
128K24.1Comfortable

Generation speed

GGUF IQ2_XXS: 786 tok/s · Bandwidth ceiling (not measured)

Other model / GPU pairs