GLM-5.3 (Low): benchmark results

Provider: Zhipu. Released 2026-08-14. Access: API.

Unified ELO 1651 ± 1, rank #217 of 1935 rated models, from 59 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Artificial Analysis Intelligence Index34.3Intelligence Index89.4
AA CritPt14.57Accuracy (%)85.9
AA Humanity's Last Exam36.56Accuracy (%)85.9
AA-Omniscience Index - Software Engineering (SWE) - TypeScript34.44Omniscience Index80.6
AA-Omniscience Index - Software Engineering (SWE) - Kotlin20Omniscience Index79.8
AA-Omniscience Index - Science, Engineering & Mathematics0.8Omniscience Index79.1
AA-Omniscience Index - Software Engineering (SWE) - C43Omniscience Index77.9
AA-Omniscience Index - Software Engineering (SWE) - Swift32Omniscience Index77.2
AA-Omniscience Index - Software Engineering (SWE) - PHP18Omniscience Index76.9
AA-Omniscience Index - Software Engineering (SWE) - R-2Omniscience Index76.2
AA-Omniscience Index - Software Engineering (SWE) - Python14Omniscience Index75.9
AutomationBench-AA - Sales - Raw Objective Completion76.88Raw Objective Completion (%)75.8

Interactive version: theaggregate.ai/model?slug=glm-5-3-low · How It Works · Data refreshed daily, snapshot 2026-09-25.