llamaperf
Submit

DGX Spark

NVIDIA · 128GB unified memory · 2 reports

See what fits on this GPU →
This page is thin (2 of 3 reports needed for indexing). Help fill it in.
LatestMost reportedFastest t/s

User is setting up a 16x DGX Spark cluster to run frontier models locally. Mentions DeepSeek V4 pro, Kimi K3, GLM 5.5, and Minimax M4 as future models. No benchmark numbers provided.

Tone: positive
quant:
NVFP4

16x DGX Spark cluster with unified memory, serving GLM-5.1-NVFP4 (434GB) at TP=8. Plans to test DeepSeek and Kimi. Future prefill/decode split with M5 Ultra Mac Studios.