DGX Spark vs RTX 5090

For running local LLMs · 15 reports across 3 models

Side A

Vendor: nvidia
VRAM: 128GB
Memory: Unified

Side B

Vendor: nvidia
VRAM: 32GB
Memory: Discrete

Tokens per second by model

Model	DGX Spark	RTX 5090
Qwen3.6up to 35B	—	215.1n=12
Gemma 4up to 31B	—	578.0n=2
GLM-5.1	—n=1	—

More comparisons