benchgap
Zhipu (GLM) · model

GLM-5.2 (max) benchmark scores

As of 2026-10-07, GLM-5.2 (max) (Zhipu (GLM)) has measured scores on 3 benchmarks and estimated scores on 2 more.

BenchmarkScoreSource
τ³-Bench Banking47.8%estimated ± 2.6 pp, medium confidence
Harvey LAB93.0%estimated ± 1.6 pp, low confidence
APEX-Agents33.7%measured
ITBench SRE42.7%measured
τ²-Bench Telecom99.1%measured