// RUN — FITS

Can the NVIDIA DGX Spark run Gemma 4 12B?

Yes — here's how fast and up to what context, with real measured numbers.

Gemma 4 12B on the NVIDIA DGX Spark

QuantFits?Max contextSpeedBeyond context
FP16 yes 547k 8.3-9.9 t/s ~0.3-0.4 (slow)
Q8 yes 621k 16.6-19.8 t/s ~0.6-0.7 (slow)
Q4 yes 653k 29.7-35.3 t/s ~1.1-1.3 (slow)

Estimate (memory-bound). Beyond "max context" the KV cache spills to SSD swap and speed drops to the "beyond" figure.

→ All models on the NVIDIA DGX Spark → Gemma 4 12B specs → Other hardware

FAQ

Can the NVIDIA DGX Spark run Gemma 4 12B?▶

Yes. At Q4 it generates ~29.7-35.3 tok/s and fits up to 653k context; beyond that it spills to SSD swap and slows to ~1.1-1.3 tok/s.

How much context fits?▶

At Q4, up to about 653k tokens stay in fast memory; longer context spills to SSD swap and throughput collapses.