Zhipu's new model ranks first globally in legal understanding

What happened: Zhipu released GLM-Legal, an open-source model for legal work such as contract review, evidence analysis and cross-border compliance research. Key numbers: 19.2% full task pass rate and 86.23% accuracy on Vals AI legal evaluations, topping the Harvey LAB and LegalBench open-weight model Why it matters: Zhipu is inviting law firms, legal tech companies, corporate legal teams and independent practitioners to apply for GLM-Legal Beta, and plans to integrate the capabilities into AutoClaw. The 19.2% pass rate means most tasks still do not fully meet strict scoring conditions.

Zhipu released GLM-Legal, an open-source model for legal work such as contract review, evidence analysis and cross-border compliance research.

Zhipu is inviting law firms, legal tech companies, corporate legal teams and independent practitioners to apply for GLM-Legal Beta, and plans to integrate the capabilities into AutoClaw. The 19.2% pass rate means most tasks still do not fully meet strict scoring conditions.