Z.ai announces GLM-Legal for professional legal workflows
Z.ai announced GLM-Legal on October 12, 2026, describing it as a system for legal work such as contract review, evidence analysis and compliance.

Key takeaways · 4
- 01
GLM-Legal targets legal workflows including contract review, evidence analysis and compliance work.
- 02
Z.ai reports a 19.2% all-pass rate on Harvey LAB, 12.5 percentage points above its base model.
- 03
The system is built on GLM-5.3-Flash and post-trained across tens of thousands of realistic legal environments.
- 04
Prospective users are asked to apply for API access and discuss collaboration; pricing and a general-availability date were not disclosed, Fathom News reported.
A legal-workflow system
Z.ai frames GLM-Legal as part of a broader shift from single-turn code generation toward professional collaboration, which it calls “Professional Cowork.”[1] The company describes legal practice as a demanding setting because work connects facts, statutory doctrine, analytical reasoning and client-ready output in a continuous chain.[1] Its stated use cases include contract review, evidence analysis and compliance work; DigitalToday also reports intended use for contract negotiation and drafting regulatory documents.[2][3] This positioning makes the announcement relevant to teams assessing whether an AI system can support connected legal tasks, rather than only isolated drafting.
Benchmark results and claims
Z.ai says it evaluated GLM-Legal on complex task execution, specialized legal reasoning and real-world legal workflows.[1] It reports a 19.2% all-pass rate on the Harvey LAB benchmark, an improvement of 12.5 percentage points over its base model.[1] The announcement also refers to real-world law-firm workflows across 24 practice areas.[1] Z.ai says Vals independently evaluated GLM-Legal’s Harvey Legal Agent Benchmark and LegalBench results.[1] These are company-reported results; the supplied information does not provide additional benchmark scores or details for comparing performance across different legal tasks.
How the system is built
GLM-Legal is built on the GLM-5.3-Flash foundation model, according to Z.ai.[1] The company says its post-training uses reinforcement learning across tens of thousands of realistic legal environments.[1] Its system manages task context, connects toolsets and routes execution feedback.[1] Z.ai further says the model parses specialized contracts, merger filings and judicial opinions with precision comparable to top closed-source frontier models despite a leaner parameter budget.[1] For a prospective evaluator, these claims point to both the model’s legal training and its orchestration of tasks and tools; they do not establish performance for every document or workflow.
Access and remaining unknowns
Z.ai’s API site asks prospective GLM-Legal users to apply for API access and discuss potential collaboration with the company’s team.[4] Fathom News reported that Z.ai had not disclosed pricing or a general-availability date.[2] DigitalToday describes GLM-Legal as additional training of GLM-5.3-Flash and reports that its operating harness was optimized for legal work.[3] Z.ai also says the model achieves parity with leading frontier systems on Chinese legal consultation, case analysis and document drafting.[1] Teams considering evaluation should clarify access terms and test the system against their own requirements before drawing conclusions from the reported results.
GLM-Legal is presented as an attempt to support linked legal tasks, from analysis to client-ready work, rather than only stand-alone text generation. Professionals evaluating it should weigh Z.ai’s benchmark and capability claims against their own workflows, while factoring in that pricing and a general-availability date were not disclosed.
Why it matters
Put this to work — one session a day, built for your industry.
Create a free account for a daily session — eight questions and one real-work challenge, on the news that affects your role.
Start freeHow this developed
14 October 2026
Z.ai announces GLM-Legal for professional legal workflows