Zhipu releases GLM-Legal, a legal-domain model
What happened
On October 13, Hacker News reported that Zhipu released GLM-Legal, a domain model aimed at agentic legal intelligence. The model builds on the lightweight GLM-5.3-Flash base and improves legal task ability through domain mid-training and efficient post-training. On Harvey LAB it reaches a 19.2% all-pass rate, 12.5 percentage points above the base model; on LegalBench and PLawBench it matches frontier closed models with a smaller parameter count.
Written by AI from the coverage · updated 57 minutes ago
Coverage
Follow the reports to see the story from different sides.
- Hacker News front pageGLM-Legal
智谱发布 GLM-Legal，基于轻量级 GLM-5.3-Flash 底座，通过领域中期训练与高效后训练提升法律任务能力。该模型在 Harvey LAB 上取得 19.2% 全通过率，较基座模型提升 12.5 个百分点，并在 LegalBench 与 PLawBench 上以更小参数规模追平前沿闭源模型。
Heat over time
Heat now 9·Comparable peak 10（Oct 13 01:00）·Comparable change over 24 hours –
The trend only compares accounts observed without gaps, so its range may be smaller than the current heat. Hover or tap the chart for each hour; the left and right arrow keys step through it.