OpenAI Introduces Astra for Law With Legal Search and Trusted Access

OpenAI on September 17, 2026, introduced Astra for Law, a configuration of its GPT-6 Astra model that pairs a legal search index spanning more than 230 million URLs with access controls and workflow tools built for law firms and legal technology companies.

OpenAI describes the offering as a new foundation for law firms and legal technology companies to build AI products and workflows around their expertise, combining GPT-6 Astra, which the company calls its latest and most powerful model, with settings, tools, and context tailored for professional legal work. API customers including Harvey and Legora will be able to build on Astra for Law, and OpenAI said it will bring the legal capabilities to its latest models as its frontier models advance.

Astra for Law will initially be offered to selected law firms through Trusted Access in ChatGPT and Codex, with API availability to follow. It will appear in the model picker as GPT-6 Astra Law and in the API as gpt-6-astra-law.

The legal configuration extends GPT-6 Astra, which OpenAI describes as its most intelligent and aligned model and which meets the Critical cybersecurity threshold under the company’s Preparedness Framework. GPT-6 Astra is rolling out to ChatGPT Plus, Pro, Business, and Enterprise users, the OpenAI API, Microsoft Azure, and AWS Bedrock.

Legal Search Index and Vals AI Benchmark Results

Using the index, Astra for Law can search U.S. case law, statutes, regulations, court rules, and administrative decisions across the corpus, with sources added daily. OpenAI said its work with Free Law Project, the nonprofit behind CourtListener, brings a case-law collection covering more than 99.9% of published U.S. precedential case law into the research experience, and that the index complements the licensed content and specialist products firms already use from providers such as Thomson Reuters.

OpenAI tested Astra for Law’s complete setup on 200 U.S. legal research questions from the private validation set of Vals AI’s Legal Research Bench, which measures source and passage retrieval and answer quality against evaluation criteria. OpenAI reported that at the highest reasoning effort for both systems, Astra for Law passed the benchmark’s overall correctness check on 54.0% of questions, versus 38.7% for GPT-6 Astra using web search alone, a result the company described as a 40% relative improvement. It also reported 24% more reference cases on case-law-focused questions and up to 54% more relevant passages retrieved from the correct court opinions on an audited set of target passages at the same reasoning effort.

Custom instructions for legal analysis and writing then guide the model in applying that research to a client’s facts, developing arguments or deal terms, and identifying weaknesses, such as distinguishing a court’s holding from its other observations or addressing cases that weaken an argument. In a worked example built on a misrepresentation-claim prompt, OpenAI said Astra for Law returned two closely matching precedents, while Claude Fable 5.1 returned a holding that had been reversed on appeal in a litigation example and reported finding no such case in a transactional example.

Trusted Access Controls and Firm-Built Tools

OpenAI created a Trusted Access Program that gives lawyers at eligible law firms, and people working under their supervision, access to Astra for Law for professional legal work. For eligible firms, the offering includes Zero Data Retention on the API, and usage of ChatGPT Enterprise is excluded from human review by default. OpenAI is also working with Latham & Watkins to design for information permissions, ethical walls, client instructions, and firm oversight. Michael Rubin, chair of Latham’s AI Strategy Committee, said the collaboration “builds on Latham’s broader investments in AI development, governance, and infrastructure across the firm, which underpin our enterprise-wide strategy for responsible AI.”

Working with selected firms, OpenAI’s forward-deployed engineers have adapted ChatGPT Enterprise with custom interfaces and integrations to proprietary data. Sullivan & Cromwell built an agreement analyzer that brings the firm’s negotiating playbooks and selected precedents into the review of a new deal, helping lawyers spot risks that emerge when provisions are read together and turning those findings into proposed redlines and draft client advice. Ropes & Gray built a deal diligence system that traces findings back to sources in a data room and flags questions that could affect an acquisition, such as whether key customer contracts require notice or consent. Cooley built GO Public, which applies the firm’s capital markets expertise to IPO preparation, from drafting the filing to carrying deal changes across the filing so lawyers can review the implications together.

John Savva, a Sullivan & Cromwell partner, said early versions of Astra for Law “demonstrated impressive research depth and sensitivity to authority” across both litigation and transactional matters. Niko Grupen, head of applied research at Harvey, said early testing showed strength in grounding answers in on-point authorities, citing with precision, and offering practical, advisory guidance. OpenAI said its work with Wachtell, Lipton, Rosen & Katz pairs the firm’s litigation and corporate expertise with OpenAI’s research and engineering to explore how AI can support the work involved in providing sophisticated legal judgment.

Partner Plugins and ChatGPT for Word

Alongside the model configuration, OpenAI launched 26 partner-built plugins that connect ChatGPT to specialist legal tools firms already use, including Relativity and Clio. Among the named integrations, iManage lets a lawyer draft a negotiation brief in ChatGPT and save it to the matter file, Intapp can surface activities that may need a time entry for review, and DeepJudge can bring prior deals into a comparison. Thomson Reuters is bringing HighQ matter context into ChatGPT and previewing a forthcoming CoCounsel Legal connector.

The launch also includes 9 community plugins from lawyers and legal engineers at LegalQuants, LECG, and Skills.law, carrying 47 custom skills that practitioners can adapt or extend for their own work in ChatGPT. OpenAI also made ChatGPT for Word generally available, letting lawyers proofread, receive suggested edits, and flag formatting issues inside Word.

OpenAI said it is investing in law for the long term and described its approach as open and composable, with firms choosing how ecosystem capabilities fit together. Astra for Law is coming soon to the API.