GLM-5.3 Open Weights: SGLang Reports Over 500 tok/s on Agentic Workloads
BanghuaZ · x · 2026-08-28
Zai has released GLM-5.3 as an open-weight model, positioning it as their most capable model for agentic coding and cyber defense. SGLang has announced day-0 support with performance benchmarks on NVIDIA Blackwell/Hopper and AMD MI300 series hardware.
Key Performance Metrics:
- On real-world multi-turn agentic workloads (BS=1, TP8):
- 537.6 tok/s/user on NVFP4
- 413 tok/s/user on FP8
The model inherits optimizations from GLM-5.2 and is production-ready on NVIDIA Blackwell, Hopper, and AMD MI300X/325X/355X.
Related event: Z.ai Open-Sources GLM-5.3 and GLM-5.3-Flash(28 posts)→
More from Infra
- Fluidstack hires Electrical Engineer for Modular R&D to build frontier AI infrastructure — MxMnr · 2026-10-16
- Jensen Huang joins BG2 podcast: does AI revenue justify capex, '27 bottlenecks — elonmusk · 2026-10-16
- Mistral's on-device infra rack revealed: DGX Spark, Mac Studio and more — gregmushen · 2026-10-16
- NVIDIA and Sweden's Wallenberg launch Sferical One, a shared AI supercomputer — nvidia · 2026-10-16
- Next infra frontier: self-evolving systems that adapt and self-verify — _shreya_s · 2026-10-16
- Xiaomi's MiMo RL dashboard reveals $3.47M spent on just 30 training steps — bycloud · 2026-10-16