ChatGLM3-6B-32K: benchmark results
Zhipu and Tsinghua KEG's ChatGLM3 long-context dialogue checkpoint supporting 32K context, distinct from the base, standard dialogue and 128K releases. Provider: Zhipu. Access: Open.
Aggregate Rating 1401 ± 23, rank #6074 of 10916 rated models, from 76 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| OpenCompass Language - Content Summarization - English (CompassBench 2404) | 35.6 | Score (%) | 96.7 |
| OpenCompass Code - HumanEval+ - English (CompassBench 2404) | 45.1 | Score (%) | 70 |
| OpenCompass Language - Content Summarization - Chinese (CompassBench 2404) | 25.4 | Score (%) | 70 |
| OpenCompass Code - HumanEval - Chinese (CompassBench 2404) | 43.3 | Score (%) | 66.7 |
| OpenCompass Code - MBPP - Chinese (CompassBench 2404) | 38.2 | Score (%) | 66.7 |
| OpenCompass Agent - T-Eval - Chinese (CompassBench 2404) | 63.6 | Score (%) | 63.3 |
| OpenCompass Code - MBPP - English (CompassBench 2404) | 54.9 | Score (%) | 63.3 |
| OpenCompass Language - Content Critique - English (CompassBench 2404) | 34 | Score (%) | 61.7 |
| OpenCompass Language - Sentiment Analysis - Chinese (CompassBench 2404) | 86.3 | Score (%) | 61.7 |
| OpenCompass Language - Translation (CompassBench 2404) | 14.6 | Score (%) | 60 |
| OpenCompass Reasoning - Common Sense - English (CompassBench 2404) | 15.2 | Score (%) | 55 |
| OpenCompass Agent - T-Eval - English (CompassBench 2404) | 62.3 | Score (%) | 53.3 |
Interactive version: theaggregate.ai/model?slug=chatglm3-6b-32k · How It Works · Data refreshed daily, snapshot 2026-11-08.