GLM-5.2 (High): benchmark results
GLM-5.2 evaluated at the high reasoning-effort setting. Provider: Zhipu. Released 2026-06-13. Access: Open.
Aggregate Rating 1745 ± 15, rank #248 of 2168 rated models, from 72 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Horangi 4 - ALT Average | 81.1 | Score (%) | 96.2 |
| Horangi 4 - Ko-HalluLens (Nonexistent Entities) | 96 | Refusal rate (%) | 96.2 |
| Horangi 4 - ALT - Hallucination Prevention | 74.46 | Score (%) | 94.2 |
| Horangi 4 - Ko-Moral | 84 | Accuracy (%) | 93.8 |
| Horangi 4 - Ko-HellaSwag | 88 | Accuracy (%) | 91.3 |
| Horangi 4 - GLP - Logical Reasoning | 87.5 | Score (%) | 89.4 |
| Horangi 4 - Final Score | 77.21 | Score (%) | 87.5 |
| Horangi 4 - IFEval-Ko | 90.42 | Instruction-following accuracy (%) | 84.6 |
| SWE-rebench | 57.01 | Resolved (%) | 83.3 |
| Horangi 4 - BFCL | 68.6 | Accuracy (%) | 81.9 |
| Horangi 4 - KMMLU | 88 | Accuracy (%) | 81.2 |
| Horangi 4 - SWE-bench Verified (80-task subset) | 56.16 | Resolved (%) | 80.8 |
Interactive version: theaggregate.ai/model?slug=glm-5-2-high · How It Works · Data refreshed daily, snapshot 2026-10-20.