OpenAI has launched Astra for Law, a new legal foundation designed for law firms and legal‑technology companies. The package pairs GPT‑6 Astra — OpenAI’s latest, most capable model — with settings, tools, and contextual instructions specifically adapted for professional legal research, drafting, and analysis.
What Astra for Law includes
Astra for Law combines GPT‑6 Astra with a dedicated legal search index and tailored instructions for legal analysis and writing. The legal index lets the system search U.S. case law, statutes, regulations, court rules, and administrative decisions across a corpus of more than 230 million URLs that is updated daily. OpenAI’s collaboration with the Free Law Project (the nonprofit behind CourtListener) brings a case‑law collection that covers more than 99.9% of published U.S. precedential case law into this research experience.
The index is intended to complement licensed content and specialist products that firms use, such as Thomson Reuters offerings.
Benchmark results on legal research
OpenAI evaluated Astra for Law using 200 U.S. legal research questions from the private validation set of Vals AI’s Legal Research Bench. That benchmark measures how well a system finds relevant sources and passages, and how well its research outputs meet evaluation criteria.
At the highest reasoning effort, Astra for Law passed the evaluation’s overall correctness check on 54.0% of questions, compared with 38.7% for GPT‑6 Astra using web search alone — a 40% relative improvement. Astra for Law also produced more comprehensive answers.
For case‑law‑focused questions, Astra for Law found 24% more reference cases than GPT‑6 Astra with web search at the same reasoning level. On an audited set of target passages, it retrieved up to 54% more relevant passages from the correct court opinions when systems were compared at the same reasoning effort.
These findings indicate a stronger research foundation for advising on transactions, assessing disputes, or developing legal strategies, with authorities that lawyers can inspect directly.
Applying research to legal tasks
Beyond retrieval, Astra for Law uses custom instructions to guide legal analysis and drafting — applying research to a client’s facts, developing arguments or deal terms, and flagging weaknesses or uncertainties. The configuration can, for example, distinguish a court’s holding from its observations, address cases that weaken an argument, or explain how a contract exception reallocates risk.
API customers including Harvey and Legora will be able to build on Astra for Law and incorporate its capabilities into their products and workflows. In ChatGPT the model will appear as “GPT‑6 Astra Law”; in the API it will be available as gpt-6-astra-law. OpenAI says it will bring these legal capabilities to future frontier models as they become available.
Privacy, governance and Trusted Access
To address confidentiality and governance needs, OpenAI is offering a Trusted Access Program for eligible law firms. For those firms the offering includes Zero Data Retention (ZDR) on the API, and ChatGPT Enterprise usage is excluded from human review by default. The company is also working with Latham & Watkins on information permissions, ethical walls, client instructions, and firm oversight.
OpenAI frames these controls as a way for firms to protect client confidences while turning their precedents, methods, and judgment into AI‑enabled tools and workflows.
Firm integrations and bespoke tools
OpenAI’s forward‑deployed engineers have been adapting ChatGPT Enterprise for selected firms with custom interfaces and integrations to proprietary data. Examples provided by OpenAI:
- Sullivan & Cromwell built an agreement analyzer that brings the firm’s negotiating playbooks and selected precedents into contract review, highlights risks across provisions, and generates proposed redlines and draft advice.
- Ropes & Gray created a deal diligence system that mirrors how lawyers work through a data room, traces findings to their sources, and surfaces questions that could affect an acquisition (for instance, notice or consent obligations in customer contracts).
- Cooley developed GO Public to support capital markets work, assisting with IPO filing drafts, surfacing risks, and propagating deal changes across documents for coordinated review.
Firms can build with their own teams and partner products, while defining permitted sources and review processes for their work.
Plugins, community skills and ecosystem partners
At launch, OpenAI is rolling out 26 partner‑built plugins that connect ChatGPT to specialist tools and systems used in legal practice. Partners and examples include iManage, Intapp, DeepJudge and Thomson Reuters (bringing HighQ matter context and previewing a forthcoming CoCounsel Legal connector). Use cases include drafting a negotiation brief in ChatGPT and saving it to a matter file in iManage, surfacing activities in Intapp that may need time entries, or comparing prior deals with DeepJudge.
The release also includes 9 community plugins from legal engineers and practitioners at LegalQuants, LECG and Skills.law, offering 47 custom skills that practitioners can adapt, extend, or use for inspiration. Additionally, ChatGPT for Word is being made generally available to support proofreading, suggested edits, and formatting checks inside a drafting tool lawyers already use.
OpenAI says its approach is open and composable: firms can continue to use specialist products, integrate their own tools and knowledge, and adapt community‑built skills rather than being forced to replace existing systems.
Ongoing development and partnerships
OpenAI states it will continue investing in Astra for Law’s models, settings, tools and instructions, guided by rigorous evaluations and feedback from lawyers and legal‑tech partners. The company cites work with Wachtell, Lipton, Rosen & Katz to explore how frontier research and engineering can support sophisticated legal judgment in litigation and corporate practice.
OpenAI invites firms and legal‑tech companies to explore early access, build Astra for Law into products, or develop tools around their firms’ expertise by contacting OpenAI.
