NVIDIA DGX Spark

shared by gpu4.pics builder

gpu4.pics

submitted

NVIDIA DGX Spark GB10 Grace-Blackwell

NVIDIA

running qwen3:14b

23.9

tokens / sec

Generation

23.9tok/s

higher is better

Prompt / prefill

—tok/s

higher is better

Time to first token

—

lower is better

Composite score

—

Model size

—

Runtime

ollama

All benchmarks on this rig

Model	Quant	Gen	Prompt	TTFT	Power	Perf/W	$/Mtok	Score
qwen3:14b	—	23.9	—	—	—	—	—	—

NVIDIA DGX Spark GB10 Grace-Blackwell

23.9

Apple M3 Max

35.3