RunsOnMy ModelsGPUs & MacsBy VRAMAPI pricesVRAM calculatorNew

Can LFM2 350M run on the A100 80GB?

Yes. At 8K context the best fit is F16, using about 1.4 GB of the 76.0 GB usable on the A100 80GB. Estimated speed ceiling: ~2876 tokens/sec.

Memory by quant and context

Quant4K ctx8K ctx32K ctxMax tok/s (est., 8K)
2-bit (Q2_K)0.8 GB Fits0.9 GB Fits1.6 GB Fits15339
3-bit (Q3_K_M)0.8 GB Fits0.9 GB Fits1.7 GB Fits11769
4-bit (Q4_K_M)0.9 GB Fits1.0 GB Fits1.7 GB Fits9410
5-bit (Q5_K_M)0.9 GB Fits1.0 GB Fits1.8 GB Fits8073
6-bit (Q6_K)0.9 GB Fits1.0 GB Fits1.8 GB Fits7015
8-bit (Q8_0)1.0 GB Fits1.1 GB Fits1.9 GB Fits5414
16-bit (F16/BF16)1.3 GB Fits1.4 GB Fits2.2 GB Fits2876

• real GGUF file size. Speed is a bandwidth ceiling (2039 GB/s), not a benchmark. Methodology.

More

LFM2 350M on every GPU and Mac · Every model that runs on the A100 80GB · Other models that fit 64 GB VRAM · Try other settings in the calculator

Data: Hugging Face Hub (LiquidAI/LFM2-350M), NVIDIA specs. Data as of 2026-11-17.