Key Notes
- OpenAI introduced Astra for Law, a new foundation combining GPT-6 Astra with a legal research index and firm-level controls for confidential client work, offered initially to selected law firms through a Trusted Access program.
- OpenAI said the configuration passed its Legal Research Bench evaluation on 54% of questions compared with 38.7% for the base model, and published a comparison in which Astra for Law outperformed Anthropic's Claude Fable 5.1 on sample legal research prompts, a claim not yet independently verified.
- The launch includes 26 partner-built plugins connecting ChatGPT to tools including iManage, Thomson Reuters and DeepJudge, along with custom tools built by firms such as Sullivan & Cromwell, Ropes & Gray and Cooley.
OpenAI has introduced Astra for Law, a new foundation that lets law firms and legal technology companies build AI products and workflows around GPT-6 Astra, the company’s latest and most powerful model, configured specifically for professional legal work.
The offering combines GPT-6 Astra with a dedicated legal search index and instructions tailored to legal analysis and writing. According to OpenAI, the index lets Astra for Law search U.S. case law, statutes, regulations, court rules and administrative decisions across a corpus of more than 230 million URLs, with sources added daily. The company said its work with Free Law Project, the nonprofit behind CourtListener, brings coverage of more than 99.9% of published U.S. precedential case law into the research experience.
Measuring the Research Gains
OpenAI tested Astra for Law’s complete setup on 200 U.S. legal research questions drawn from the private validation set of Vals AI’s Legal Research Bench. At the highest reasoning effort for both systems, the company said Astra for Law passed the benchmark’s overall correctness check on 54% of questions, compared with 38.7% for GPT-6 Astra using web search alone, a 40% relative improvement.
On case-law-focused questions, OpenAI said Astra for Law found 24% more reference cases and retrieved up to 54% more relevant passages from the correct court opinions than the base model. These figures come from OpenAI’s own published benchmark and have not yet been independently verified.
OpenAI’s announcement also included a head-to-head comparison against a rival model, Anthropic’s Claude Fable 5.1, on sample litigation and transactional research prompts.
In OpenAI’s example, Astra for Law returned two closely matching precedents while Claude Fable 5.1 cited a holding that had been overturned on appeal in one case and reported finding no matching case in the other. Anthropic has not issued a public response to the comparison, and independent legal researchers have not yet reviewed the claim.
Access, Controls and Early Partners
Astra for Law will initially be offered to selected law firms through a Trusted Access program in ChatGPT and Codex, appearing in the model picker as GPT-6 Astra Law, with API access coming soon under the identifier gpt-6-astra-law. API customers including Harvey and Legora will be able to build the model into their own products.
OpenAI said eligible firms get Zero Data Retention on its API, and that ChatGPT Enterprise usage is excluded from human review by default. The company is also working with Latham & Watkins to design permissions, ethical walls and firm-level oversight controls for confidential client work.
Several firms have built custom tools on top of the platform through OpenAI’s forward-deployed engineering teams. Sullivan & Cromwell built an agreement analyzer that folds the firm’s negotiating playbooks into deal review, Ropes & Gray built a diligence system for working through data rooms, and Cooley built a tool called GO Public to support IPO preparation. Executives from all three firms offered supportive quotes in OpenAI’s announcement.
Part of a Broader Legal AI Ecosystem
Alongside Astra for Law, OpenAI launched 26 partner-built plugins connecting ChatGPT to tools firms already use, including iManage, Intapp, DeepJudge and Thomson Reuters’ HighQ, along with 9 community plugins and 47 custom skills from independent legal engineers. The company also made ChatGPT for Word generally available, letting lawyers proofread and flag formatting issues inside the word processor they already use for drafting.
The launch adds legal services to a growing list of industries where OpenAI is positioning ChatGPT and its underlying models as an enterprise platform rather than a general-purpose chatbot, following moves such as the shifting enterprise preferences among corporate AI buyers and OpenAI’s continued rollout of specialized capabilities on top of GPT-6 Astra since the model’s release.
OpenAI said it will keep advancing Astra for Law’s model, settings and tooling based on feedback from lawyers and legal technology partners, and that firms interested in early access can contact the company directly.
OpenAI described its approach as open and composable, saying firms can use specialist products, bring their own tools and knowledge, and adapt community-built skills rather than being locked into a single vendor stack.
The company also disclosed a separate collaboration with Wachtell, Lipton, Rosen & Katz to explore how AI can support the judgment-heavy work involved in complex litigation and corporate matters, alongside the productized tools already built with Sullivan & Cromwell, Ropes & Gray and Cooley.
Disclaimer: AIstify is an independent media brand owned and operated by NuvexMedia LLC, publishing news, research, and insights on artificial intelligence, emerging technologies, automation, and related industries. NuvexMedia LLC invests in and collaborates with companies across the AI, technology, software, and digital innovation sectors. These relationships do not influence AIstify’s editorial coverage, and the publication maintains full editorial independence to provide accurate, timely, and objective information. © 2026 NuvexMedia LLC. All rights reserved. This content is for informational purposes only and should not be considered legal, tax, investment, financial, or other professional advice.
