Z.aiLaunchAI ModelsAugust 18, 2026

Z.ai's GLM-5.3 tops open-weight models, draws cyber warnings

GLM-5.3 scores 60 on the Artificial Analysis Intelligence Index, tying Kimi K3 and up 7 points from GLM-5.2, and hits 84.5% on the CyberGym vulnerability benchmark. It uses the same base model as GLM-5.2, with gains from post-training; weights release two weeks after launch.

People · Greg Brockman

How this story unfolded

3 weeks · 8 reports · 17 community posts · 25 of 26 shown

  1. Aug 14
  2. Aug 15
  3. Aug 16
  4. Aug 17
  5. Aug 18
  6. Aug 23
  7. Aug 24
  8. Aug 28
  9. Aug 31
  10. Sep 1

More stories today

Open the live feed