GLM-5.3 (Max): benchmark results
Provider: Zhipu. Released 2026-08-14. Access: Open.
Aggregate Rating 1824 ± 8, rank #96 of 2168 rated models, from 299 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SuperCLUE General (July 2026) - Agentic Task Planning | 91.15 | Score | 100 |
| Chatbot Arena (Text - German) | 1514 | Arena Score | 99.4 |
| Nejumi 4 - jaster (2-shot) - JSeM | 84 | Exact match (%) | 99 |
| Nejumi 4 - Toxicity - Fairness | 100 | Criteria met (%) | 98.6 |
| Tau3 Banking | 50.31 | Success Rate (%) | 98.6 |
| Chatbot Arena (Text - Korean) | 1477 | Arena Score | 97.9 |
| Chatbot Arena (Text - French) | 1514 | Arena Score | 97.4 |
| Nejumi 4 - jaster (2-shot) - MMLU-ProX (Japanese) | 91 | Exact match (%) | 97.2 |
| AutomationBench-AA - Finance - Raw Objective Completion | 91.17 | Raw Objective Completion (%) | 96.9 |
| Nejumi 4 - MT-Bench (Japanese) - Roleplay | 99.5 | Judge rating (1-10, x10) | 96.9 |
| Chatbot Arena (Text - Chinese) | 1535 | Arena Score | 96.3 |
| Nejumi 4 - jaster (0-shot) - JCommonsenseQA | 100 | Exact match (%) | 96.2 |
Interactive version: theaggregate.ai/model?slug=glm-5-3-max · How It Works · Data refreshed daily, snapshot 2026-10-20.