GLM-4 9B Chat: benchmark results
Zhipu AI's open GLM-4 9B chat model, the small tier of the GLM-4 family (June 2024). Provider: Zhipu. Released 2024-06-05. Access: Open.
Aggregate Rating 1464 ± 9, rank #916 of 1656 rated models, from 388 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass Language - Dialogue - English (CompassBench 2407) | 55.5 | Score (%) | 91.2 |
| Fin-Bias | 98.1 | Average Herding Score (with rating) (self-reported) | 88.9 |
| Open PL LLM - PSC (generative, 5-shot) | 97.42 | Binary F1 (%) | 85 |
| RVDB | 4.89 | Average decision flip rate (%) over the ten pairwise role-ge | 83.3 |
| OpenCompass Language - Dialogue - English (CompassBench 2409) | 48.6 | Score (%) | 82.8 |
| Open PL LLM - EQ-Bench | 53.75 | EQ-Bench Score | 82.6 |
| Open PL LLM - PolQA Reranking (multiple choice, 5-shot) | 79.96 | Accuracy (%) | 82.1 |
| Open PL LLM - DYK (generative, 5-shot) | 66.98 | Binary F1 (%) | 80.5 |
| Open PL LLM - DYK (multiple choice, 5-shot) | 66.34 | Binary F1 (%) | 79.8 |
| Open PL LLM - RAG | 69.3 | Average RAG Score (%) | 76.9 |
| Open PL LLM - PolEmo2-IN (multiple choice, 5-shot) | 82.55 | Accuracy (%) | 75.1 |
| Open PL LLM - PoQuAD Open Book (generative, 5-shot) | 68.16 | Levenshtein Similarity (%) | 74.8 |
Interactive version: theaggregate.ai/model?slug=glm-4-9b-chat · How It Works · Data refreshed daily, snapshot 2026-10-20.