{"slug": "openai-releases-astra-for-law-a-gpt-6-model-tailored-for-legal-work-targeting", "title": "OpenAI Releases Astra for Law, A GPT-6 Model Tailored for Legal Work, Targeting Large Firms and Tech Vendors", "summary": "OpenAI released Astra for Law, a configuration of its GPT-6 Astra model tailored for legal work, featuring a dedicated Legal Search Index spanning more than 230 million URLs of U.S. case law, statutes, regulations, court rules and administrative decisions drawn from CourtListener, the Free Law Project's legal research library. The product will be available soon to API customers including Harvey and Legora and, initially, to selected Am Law 200 firms through a new Trusted Access program that includes zero data retention for the API and exclusion of ChatGPT Enterprise from human review by anyone at OpenAI. OpenAI also launched 26 legal plugins from vendors such as Thomson Reuters, Harvey, Legora and iManage, nine built by lawyers and legal engineers, 47 custom skills, and general availability of ChatGPT for Word, following its June hiring of Ironclad cofounder and former CEO Jason Boehmig to lead legal product development and Anthropic's earlier launch of Claude for Legal.", "body_md": "Calling it the beginning of its long-term investment in law, ChatGPT developer [OpenAI](https://openai.com/) today released Astra for Law, a version of its latest model, [GPT-6 Astra](https://openai.com/index/gpt-6-astra/), configured for legal work, including a dedicated legal search index and custom instructions for legal analysis and writing.\n\n“Today, we’re introducing Astra for Law: a new foundation for law firms and legal technology companies to build AI products and workflows around their expertise,” the company said in a [blog post announcing the product](https://openai.com/index/astra-for-law/). “It combines GPT-6 Astra, our latest and most powerful model, with settings, tools, and context tailored for professional legal work — from researching complex cases to structuring consequential deals.”\n\nThe company said that Astra for Law will be available soon to API customers, including Harvey and Legora, and offered, at least initially, to selected law firms through a new “Trusted Access” program targeting Am Law 200 firms that grants specialized access through ChatGPT and Codex, its software engineering agent.\n\nThe Trusted Access program includes a new zero data retention policy for the API and exclusion of ChatGPT Enterprise from human review by anyone at OpenAI.\n\nFirms and legal technology companies with access to Astra for Law will be able to build on its foundation without having to assemble and maintain the underlying legal search and instructions on their own, the company said.\n\nIn addition to launching Astra for Law, OpenAI said it is launching a series of new plugins designed to make ChatGPT more useful and extensible for legal work, including 26 from vendors such as Thomson Reuters, Harvey, Legora and iManage and nine built by lawyers and legal engineers, as well as 47 custom skills.\n\nAlso, OpenAI announced the general availability of ChatGPT for Word, enabling lawyers to use ChatGPT within Word to proofread, get suggested edits, and flag formatting issues.\n\nAll of this follows the company’s [hiring in June of Jason Boehmig](https://www.lawnext.com/2026/06/ironclad-founder-jason-boehmig-joins-openai-to-develop-products-for-the-legal-sector.html), the cofounder and former CEO of contract lifecycle management company Ironclad, to lead development of products for the legal industry.\n\nAhead of that hiring, OpenAI competitor Anthropic had announced its own major push into legal with the [launch of Claude for Legal](https://www.lawnext.com/2026/05/anthropic-goes-all-in-on-legal-releasing-more-than-20-connectors-and-12-practice-area-plugins-for-claude.html).\n\n*Related: [On LawNext: Inside Claude for Legal — Anthropic’s Mark Pike on AI’s Next Frontier in Law](https://www.lawnext.com/2026/07/on-lawnext-inside-claude-for-legal-anthropics-mark-pike-on-ais-next-frontier-in-law.html).* \n\n### **Embedded Legal Research**\n\nYesterday, Boehmig and [Sherwin Wu](https://www.linkedin.com/in/sherwinwu1/), the OpenAI engineer who has led the development of the legal product, provided a briefing for the media on Astra for Law.\n\nBoehmig emphasized that this is not a new model, but rather a model configuration to customize the model for legal work. It includes domain-specific instructions on tasks such as legal writing, settings such as the length of responses, and tooling specific to the legal industry, all of which will grow over time.\n\nThe first major implementation of that tooling is its Legal Search Index, a legal research capability that allows it to search U.S. case law, statutes, regulations, court rules and administrative decisions across a corpus of more than 230 million URLs.\n\nTo create its collection of legal research materials, OpenAI drew from CourtListener, the legal research library maintained by the nonprofit Free Law Project.\n\nBeyond simply conducting legal research, OpenAI says that Astra for Law is designed to use that Legal Search Index to help lawyers apply the law to the facts of a matter, to develop arguments or deal terms, and to identify weaknesses or uncertainty.\n\nDuring the briefing, Wu demonstrated three tasks, in which he used Astra for Law in ChatGPT to draft a motion to dismiss for a hospital defending a disability-discrimination and retaliation claim, a stockholder challenge to an asset sale rushed through over a holiday weekend, and for a sales-commitment dispute.\n\n### Product Benchmarking\n\nOpenAI says it tested Astra for Law’s legal research capabilities against 200 U.S. legal research questions from the private validation set of [Vals AI’s Legal Research Bench](https://vals.ai), a benchmark that measures how well the model can find relevant sources and passages, and how well its research answers meet the evaluation criteria.\n\nAt the highest reasoning effort for both systems, the company says, Astra for Law passed the evaluation’s overall correctness check on 54% of questions, compared with 38.7% for GPT-6 Astra using web search alone – an increase of 15.3 percentage points, or a 40% relative improvement.\n\nAstra for Law also produces more comprehensive answers, roughly twice as long on average as the Astra web-only baseline across the tested reasoning settings, the company says.\n\nOn case-law-focused questions, Astra for Law found 24% more reference cases than GPT-6 Astra using web search alone at the highest reasoning effort. On the audited set of target passages, it retrieved up to 54% more relevant passages from the correct court opinions, when comparing the systems at the same reasoning effort.\n\n### **New Partner and Community Plugins**\n\nAs part of today’s announcement, OpenAI is also launching a series of vendor- and community-developed plugins designed to help legal professionals do more with ChatGPT, including 26 developed by partners such as Thomson Reuters, Harvey, Legora and iManage.\n\nAs examples of what the vendor plugins will enable, OpenAI said iManage’s will allow a lawyer to draft a negotiation brief in ChatGPT and save it to the matter file, Intapp’s will surface activities that may need a time entry for review, DeepJudge’s will bring prior deals into a comparison, and Thomson Reuters’ will bring HighQ matter context into ChatGPT and later include a connector to CoCounsel Legal.\n\nThere are also nine community plugins developed by lawyers and legal engineers at [LegalQuants](https://platform.legalquants.com/), [LegalEng Consulting Group](https://www.lecg.com/) (LECG), and [Skills.law](https://skills.law/), and 47 custom skills that legal professionals can adapt, extend or draw inspiration from to do work in ChatGPT. OpenAI said these custom skills were developed by legal “power users,” not by OpenAI.\n\n### **New ‘Trusted Access Program’**\n\nAs part of the new offering, the company is creating a special Trusted Access Program under which eligible law firms will be able to give lawyers and people working under their supervision access to Astra for Law for professional legal work.\n\nThe program is targeted at Am Law 200 programs. For firms that qualify, the offering includes zero data retention on the API and usage of ChatGPT enterprise is excluded from human review.\n\nIn addition, OpenAI said it is working with the law firm Latham & Watkins to design for information permissions, ethical walls, client instructions, and firm oversight.\n\nAsked by a reporter whether there were any exceptions to the zero data retention policy, Boehmig said that is a one-sentence summary of the full access-program agreement, which runs some 30 pages. But he said OpenAI has worked with the industry to validate that it works in this context. “It’s definitely more complicated than the one sentence,” he said, “but we also feel extremely confident that the 30 pages do meet the need.”\n\n### **Building On Firms’ Own Knowledge**\n\nOpenAI’s announcement stresses the importance for law firms of building AI around their own precedents, methods and judgment, and offers several examples of how its forward-deployed engineers — OpenAI engineers who are embedded within a law firm — have been adapting ChatGPT Enterprise through custom integrations with firms’ proprietary data and workflows.\n\nIt says that Sullivan & Cromwell, for example, built an agreement analyzer that uses the firm’s negotiating playbooks and selected precedents in the review of a new deal. Another firm, Ropes & Gray, built an M&A diligence system around how its lawyers work through a data room and decide what matters to the deal.\n\nStill another, Cooley, built GO Public to bring its capital markets expertise into how companies prepare to go public, from drafting the IPO filing to identifying the risks that deserve management’s attention.\n\n“With frontier intelligence and the right controls, firms can turn their own precedents, methods, and judgment into AI tools and workflows built to their standards,” OpenAI says.\n\n### **Just the Beginning**\n\n“OpenAI is investing in law for the long term,” the company said in making these announcements.\n\nIt said that it will continue to improve ChatGPT as a platform for firms to build around their expertise and connect the tools they already use, and to conduct rigorous evaluations and collect user feedback to monitor where improvements are needed and what to bulid next.\n\nIn response to a question at the media briefing, Boehmig said the company’s commitment to legal includes a commitment to access to justice, which he said is a “paramount concern” to him personally and to OpenAI as a company, but that today’s news is focused on Astra for Law and that there will be more to share in the future related to A2J.\n\n“We’re committed to working with this ecosystem so firms can bring their own expertise, standards, and judgment to AI – and serve clients in ways only they can,” the company said.", "url": "https://wpnews.pro/news/openai-releases-astra-for-law-a-gpt-6-model-tailored-for-legal-work-targeting", "canonical_source": "https://www.lawnext.com/2026/09/openai-releases-astra-for-law-a-gpt-6-model-configured-for-legal-work.html", "published_at": "2026-09-17 20:00:31+00:00", "updated_at": "2026-09-17 20:25:10.035321+00:00", "lang": "en", "topics": ["ai-products", "large-language-models", "ai-tools", "generative-ai"], "entities": ["OpenAI", "Astra for Law", "GPT-6 Astra", "Jason Boehmig", "Sherwin Wu", "Harvey", "Legora", "Anthropic"], "alternates": {"html": "https://wpnews.pro/news/openai-releases-astra-for-law-a-gpt-6-model-tailored-for-legal-work-targeting", "markdown": "https://wpnews.pro/news/openai-releases-astra-for-law-a-gpt-6-model-tailored-for-legal-work-targeting.md", "text": "https://wpnews.pro/news/openai-releases-astra-for-law-a-gpt-6-model-tailored-for-legal-work-targeting.txt", "jsonld": "https://wpnews.pro/news/openai-releases-astra-for-law-a-gpt-6-model-tailored-for-legal-work-targeting.jsonld"}}