GLM 5.2 Speeds Up on DGX Spark
WesEklund · x · 2026-07-15
Shared details revealed an inference optimization test for GLM 5.2: after running on 2 DGX Spark units, the speed exceeded 18 tok/s.
The author explained that a key modification was made: quantizing the 3 largest attention projections to NVFP4, which still showed room for improvement compared to the original FP8. They also added that there is further room for optimization on both the speed and quality fronts.
More from Infra
- Nearly half of Claude revenue now flows through cloud marketplace commitments, up from 11% — BenBajarin · 2026-10-26
- Anthropic ARR goes vertical after price cuts: tokens up 4x, $100B+ ARR back in play — GavinSBaker · 2026-10-26
- Alex Banks argues power caps GPU supply, real risk is demand cooling first — thealexbanks · 2026-10-26
- Analyst: Ant has the most fungible AI workload of any model provider today — BenBajarin · 2026-10-26
- Anthropic S-1: training infra edge cut per-task cost 40% and doubled revenue growth — BenBajarin · 2026-10-26
- Rewriting scaling law study: optimal synthetic data share rises from 10% to 32% with compute — heghbalz · 2026-10-26