GLM-5.3

Fri, 14 Aug 2026 00:00:00 +0000

Z.ai team has released GLM-5.3 which performs as good as proprietary models like Fable, GPT-5.6 Sol, etc

And it is built on top of GLM-5.2

Scaling post-training is all we did for GLM-5.3

What’s new?

GLM-5.3 supports three thinking effort levels: low, high, and max. Disabling thinking is no longer supported by GLM-5.3.

It’s available on GLM Coding Plan & ZCode as well.