GLM-Zero-Preview: benchmark results
Provider: Zhipu. Released 2024-12-31. Access: API.
Unified ELO 1579 ± 43, rank #712 of 2656 rated models, from 63 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass LLM - Math - Chinese (CompassBench 2501) | 42 | Score (%) | 89.7 |
| OpenCompass Math - Competition - Chinese (CompassBench 2501) | 30 | Score (%) | 89.7 |
| OpenCompass Math - Reasoning Computation - Chinese (CompassBench 2501) | 31.9 | Score (%) | 88.5 |
| OpenCompass LLM - Math (CompassBench 2501) | 42 | Score (%) | 87.2 |
| OpenCompass Reasoning - Common - Chinese (CompassBench 2501) | 74.3 | Score (%) | 87.2 |
| OpenCompass Reasoning - Puzzle - English (CompassBench 2501) | 83.2 | Score (%) | 87.2 |
| OpenCompass LLM - Math - English (CompassBench 2501) | 42 | Score (%) | 84.6 |
| OpenCompass LLM - Reasoning - English (CompassBench 2501) | 73 | Score (%) | 84.6 |
| OpenCompass Math - College - English (CompassBench 2501) | 68 | Score (%) | 84.6 |
| OpenCompass Math - Reasoning Computation - English (CompassBench 2501) | 38.1 | Score (%) | 84.6 |
| OpenCompass Reasoning - Common - English (CompassBench 2501) | 77 | Score (%) | 84.6 |
| SuperCLUE General (March 2025) - Science Reasoning | 64 | Score | 84.3 |
Interactive version: theaggregate.ai/model?slug=glm-zero-preview · How It Works · Data refreshed daily, snapshot 2026-09-19.