GLM-5.2 (Non-reasoning) — benchmark results
Provider: Zhipu. Released 2026-06-13. Access: Open.
Unified ELO 1717 ± 27, rank #243 of 1746 rated models, from 35 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| UGI - Writing | 61.21 | Writing Score | 94.2 |
| UGI - Natural Intelligence | 54.88 | NatInt Score | 92.6 |
| AA GDPval | 1387.57 | ELO | 85 |
| GDPval-AA | 1388 | Elo | 84.2 |
| AA Omniscience - Software Engineering (SWE) - PHP | 48 | Accuracy (%) | 83.1 |
| AA CritPt | 3.14 | Accuracy (%) | 82.9 |
| AA Omniscience | -6.4 | Score | 82.8 |
| Artificial Analysis Intelligence Index | 34.05 | Intelligence Index | 82.8 |
| AA Omniscience - Software Engineering (SWE) - HTML | 46 | Accuracy (%) | 75.6 |
| AA Omniscience - Software Engineering (SWE) - Python | 33 | Accuracy (%) | 74.9 |
| AA Omniscience - Software Engineering (SWE) - Go | 28 | Accuracy (%) | 72.4 |
| Tau3 Banking | 15.67 | Success Rate (%) | 69.8 |
Interactive version: aibenchmarks.dev/model?slug=glm-5-2-non-reasoning · How the rankings work · Data refreshed daily, snapshot 2026-07-11.