Z.ai GLM-5.3 Coding Metrics Surge 50% Over Predecessor

Z.ai releases GLM-5.3 with 50% coding boost and top open-source rankings; weights available in two weeks.

Woofun AI reports that Z.ai has released GLM-5.3, a model built on the same foundation as GLM-5.2 but enhanced through expanded training. The update delivers a 50% improvement over GLM-5.2 on internal Code Bench metrics and ranks at the forefront among open-source models on Terminal Bench 3 and Agents' Last Exam.

GLM-5.3 achieved an 84.5% score in CyberGym vulnerability detection tests, showing significant gains in chain exploitation benchmarks like ExploitBench and ExploitGym.