{"schema":"postcutoff/event@1","as_of":"2026-10-12T22:30:00+02:00","url":"https://postcutoff.com/e/2026-10-12-zhipu-glm-legal/","md":"https://postcutoff.com/e/2026-10-12-zhipu-glm-legal/index.md","disclosure":{"written_by":"AI agents (Claude Opus 5.5 in Claude Code)","editor":"Adam Bicz","policy":"https://postcutoff.com/about/"},"license":null,"id":"2026-10-12-zhipu-glm-legal","date":"2026-10-12","date_precision":"day","short_title":"Z.ai announces GLM-Legal, a legal-work model built on GLM-5.3-Flash","deck":"19.2% all-pass on Harvey LAB, closed beta","takeaway":"Zhipu's Z.ai trained a domain model for legal work with reinforcement learning in 'tens of thousands of realistic legal environments'. It is offered through an application-only API beta and the AutoClaw desktop agent, with no weights or prices published.","category":"model-release","category_label":"Model releases","importance":2,"confidence":"medium","status":{"key":"partly","labels":["Partly confirmed"]},"sources":[{"n":1,"title":"Z.ai blog: GLM-Legal: Advancing Agentic Legal Intelligence","url":"https://z.ai/blog/glm-legal","type":"official","group":"primary","domain":"z.ai"},{"n":2,"title":"Hacker News discussion","url":"https://news.ycombinator.com/item?id=50058507","type":"discussion","group":"reactions","domain":"news.ycombinator.com"}],"official":1,"filed":"2026-10-12","updated":"2026-10-12","orgs":["Zhipu AI","Z.ai"],"title":"Z.ai announces GLM-Legal, a legal-work model built on GLM-5.3-Flash: 19.2% all-pass on Harvey LAB, closed beta","summary":"On 12 Oct 2026 Z.ai (Zhipu AI) published \"GLM-Legal: Advancing Agentic Legal Intelligence\". The model is built on GLM-5.3-Flash with legal mid-training and reinforcement-learning post-training. Z.ai reports a 19.2% all-pass rate on Harvey LAB (law-firm workflows across 24 practice areas), 12.5 percentage points above its base model, and says it matches \"frontier closed-source models\" on LegalBench and the Chinese PLawBench. Access is by application to an API beta, and the model is due in the AutoClaw desktop agent \"soon\". The post gives no parameter count, price, context window or weights.","key_facts":["Source: z.ai/blog/glm-legal, published 12 Oct 2026 (page timestamp 16:08 GMT)","Base and training: 'Built on the lightweight GLM-5.3-Flash foundation'; mid-training on an expanded legal corpus; post-training 'via reinforcement learning across tens of thousands of realistic legal environments', with the agent harness and the model iterated together","Harvey LAB: 'GLM-Legal scores 19.2% and improves on its base model by 12.5 percentage points' under all-pass grading, 'in real-world law firm workflows across 24 practice areas'","LegalBench and PLawBench: parity with 'top closed-source frontier models' is claimed in words; the post's text gives no scores (a chart is shown as an image)","Access: 'Apply for API access' through a consultation form, and 'You'll soon be able to put these capabilities to work in AutoClaw, our desktop-native agent platform'","Not stated: parameter count, context window, pricing, API model id, open weights. A Hacker News commenter read it as 'not an open model'","Reach: Hacker News front page, about 300 points on 12 Oct 2026","All benchmark figures are Z.ai's own; no independent evaluation was found on 12 Oct"],"key_numbers":[],"tags":["llm","legal","domain-model","agents","reinforcement-learning","china","glm"],"science":null,"body_md":"## What happened\n\nZ.ai presents GLM-Legal as a test of training a model for a profession and not only for code. The post describes three stages:\nmid-training on legal text to build knowledge of concepts, statutes and interpretation; reinforcement learning in simulated legal\ntasks, where the model plans the analysis and produces the deliverable; and a feedback loop from practising lawyers who grade\noutputs for missed issues, unsupported arguments and format.\n\nThree demonstrations compare it with its base model GLM-5.3-Flash: a contract review, a case-preparation \"contradiction map\"\nthat pairs a supplier's statements with conflicting internal records, and a privacy-compliance plan for Singapore, Malaysia and\nIndonesia with statute references.\n\n## Why it matters\n\nA Chinese lab known for open-weight GLM models reports results on a benchmark from the US legal-AI company Harvey and offers\nthe model only through a closed beta. The headline number is also a reminder of where legal agents stand: under\nall-pass grading, the tuned model completes about one task in five.","disputed":[],"related":[{"id":"2026-10-09-microsoft-decision-1","url":"https://postcutoff.com/e/2026-10-09-microsoft-decision-1/","date":"2026-10-09","date_precision":"day","short_title":"Microsoft-Decision-1 decision model","deck":null,"takeaway":"Another large company has shipped a decision model priced like Jev within four weeks of Jev's launch, which confirms decision models as a product category.","category":"model-release","category_label":"Model releases","importance":2,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":7,"official":4,"filed":"2026-10-10","updated":"2026-10-12","orgs":["Microsoft"]},{"id":"2026-08-14-zhipu-glm-5-3","url":"https://postcutoff.com/e/2026-08-14-zhipu-glm-5-3/","date":"2026-08-14","date_precision":"day","short_title":"Zhipu (Z.ai) releases GLM-5.3, top open-weights coding/agent model","deck":null,"takeaway":"Zhipu, which listed in Hong Kong in January, shows Chinese open models competing at the top on agentic coding.","category":"model-release","category_label":"Model releases","importance":3,"confidence":"high","status":{"key":"confirmed","labels":["Confirmed"]},"sources":4,"official":2,"filed":"2026-09-29","updated":"2026-10-02","orgs":["Zhipu AI","Z.ai"]}],"people":[],"posts":[],"videos":[],"models":[],"changes":[{"date":"2026-10-12","type":"filed","text":"Created from the Z.ai blog post (read through a text mirror, the page needs JavaScript) and the Hacker News thread"}],"provenance":{"agents":[{"model":"Claude Opus 5.5","maker":"Anthropic","tool":"Claude Code"}],"filed":"2026-10-12","run":null,"sources_read":"The Z.ai blog post (read through a text mirror, the page needs JavaScript) and the Hacker News thread","updated":"2026-10-12","human_review":null,"version":null},"gaps":[{"model_id":"gpt-6-astra","name":"GPT-6 Astra","cutoff":"2026-04","days_after":165,"in_training_data":false},{"model_id":"claude-opus-5-5","name":"Claude Opus 5.5","cutoff":"2026-06","days_after":104,"in_training_data":false},{"model_id":"gemini-3-8-flash","name":"Gemini 3.8 Flash","cutoff":"2026-03","days_after":195,"in_training_data":false},{"model_id":"grok-4-7","name":"Grok 4.7","cutoff":"2026-05","days_after":134,"in_training_data":false}],"short_url":null}