OpenAI Group PBC launched Astra for Law, a version of its GPT-6 Astra model configured for legal research and drafting. The company did not train a new model: it wrapped GPT-6 Astra in a legal search index and a set of instructions for legal analysis, then offered the result to law firms and to the software companies that sell to them. The benchmark OpenAI chose rewards finding the right authority and the right passage inside it, and on that measure the configuration passed the overall correctness check on 54% of 200 questions, against 38.7% for GPT-6 Astra with web search alone.
What is inside the legal index
The index reaches U. S. case law, statutes, regulations, court rules and administrative decisions across more than 230 million URLs, with sources added daily. The nonprofit Free Law Project's CourtListener database supplies the case law. Both systems were run at their highest reasoning effort on the same 200 questions, drawn from the private validation set of Vals AI's Legal Research Bench. On case-law questions Astra for Law found 24% more reference cases, and on a separate audited set it retrieved up to 54% more of the target passages from the correct opinions. Those two figures describe different tasks: locating the relevant case, and locating the exact passage inside it.
Access is limited at the start. Selected firms reach the configuration through a Trusted Access program in ChatGPT and Codex, where it appears in the model picker as GPT-6 Astra Law. An API version, gpt-6-astra-law, is due later, with no date or pricing attached; legal AI companies Harvey Technologies Inc. and Legora AB are named as API customers for it. For eligible firms the offering carries zero data retention on the API, and ChatGPT Enterprise usage is excluded from human review by default. Latham & Watkins is working with OpenAI on governance design for information permissions, ethical walls and client instructions.
Twenty-six partner-built plugins launched alongside the model, connecting ChatGPT to tools such as Relativity, Clio, iManage and DeepJudge. Thomson Reuters Corp. is bringing HighQ matter context into ChatGPT and previewing a connector for its CoCounsel Legal product. An additional nine plugins came from lawyers and legal engineers outside the vendor set, carrying 47 custom skills. ChatGPT for Word reached general availability the same day. Joel Hron, chief technology officer at Thomson Reuters, said the link between ChatGPT and legal tools is only part of what firms need: as AI becomes more open and interoperable, the value is not in connectivity alone, and legal professionals need trusted intelligence, relevant enterprise and matter context, purpose-built legal capabilities and governance for high-stakes work.
What this means for business
For law firms and in-house teams the practical change is that a general model is now sold with a curated corpus and a defined workflow attached, rather than as a raw tool that each firm must ground in its own sources. A small practice without a data engineering team gets a working research setup through ChatGPT, while a large firm can push the same model through the API into its own systems and keep its document handling rules. The nine outside plugins carrying 47 custom skills show that part of the configuration is being built by practitioners rather than by the vendor, which shortens the distance between a released feature and a usable one.
What remains unverified is how the benchmark numbers hold up on a firm's own matters. The 54% correctness figure comes from 200 questions in a private validation set, not from production research, and the API has no date or pricing, so cost per seat cannot be modeled yet. Zero data retention applies to the API for eligible firms, while ChatGPT Enterprise usage is excluded from human review by default — the two are separate commitments and should be confirmed in writing. OpenAI engineers embedded at individual firms have also been building firm-specific tools on ChatGPT Enterprise: Sullivan & Cromwell has an agreement analyzer that pulls its negotiating playbooks and selected precedents into the review of a new deal and turns what it finds into proposed redlines, Ropes & Gray worked on deal diligence, and Cooley's tool GO Public handles initial public offering preparation, drafting the filing included. Those builds are not the same product as Astra for Law, and a buyer should ask which one is being demonstrated.
The marker to watch is the API release of gpt-6-astra-law together with its pricing, because that is the point at which legal software vendors can embed the configuration in their own products and the offering stops being a limited program for selected firms. OpenAI described the launch as the start of a long-term investment in law, and GPT-6 Astra itself started rolling out two weeks earlier, on Sept. 3. If the API arrives with terms that let Harvey, Legora and their peers ship it at scale, legal research moves from a specialist purchase to a component inside tools firms already use.
