Skip to content

Can A100 40G run Kolibri-1 78B-A3.5B?

Comfortable: GGUF IQ2_XS needs 25.6 GB at 4K, batch 1; headroom 14.4 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 IQ1_M20.9ComfortableVerified file
GGUF IQ2_XXS24.0ComfortableVerified file
GGUF IQ2_XS25.6ComfortableVerified file
GGUF IQ3_XXS35.4TightVerified file
GGUF Q3_K_M39.5TightVerified file
GGUF IQ4_XS45.8Does not fitVerified file
GGUF Q4_K_M52.2Does not fitVerified file
GGUF Q8_087.1Does not fitVerified file

Context capacity · GGUF IQ2_XS

ContextTotal GBVerdict
4K25.6Comfortable
16K25.9Comfortable
32K26.2Comfortable
128K28.3Comfortable

Generation speed

GGUF IQ2_XS: 1503 tok/s · Bandwidth ceiling (not measured)

Other model / GPU pairs