llamaperf

DGX Spark vs RX 7900 XTX

For running local LLMs · 37 reports across 10 models

Side A
DGX Spark
Vendor
nvidia
VRAM
128GB
Memory
Unified
Side B
RX 7900 XTX
Vendor
amd
VRAM
24GB
Memory
Discrete

Highest reported generation speed by model family

These maxima come from different community setups. Model size, quantization, GPU count, offloading and concurrency may differ. They do not establish which GPU is faster for the same workload. Check the comparison conditions →

ModelDGX SparkRX 7900 XTX
Qwen3.8up to 2400B181.0n=472.1n=8
DeepSeek V4 Flashup to 285B90.0n=1012.0n=1
Ling-3.0up to 124B40.9n=4
GLM-5.2up to 744B24.0n=2
Gemma 4up to 31B30.1n=2
Qwen3.6up to 35Bn=1n=1
DeepSeek V3n=1
GLM-5.1n=1
Kimi K320.0n=1
Museup to 30B37.0n=1

More comparisons