OpenAI has introduced Astra for Law, a version of GPT-6 Astra tailored specifically for professional legal work. The system combines OpenAI’s latest model with a dedicated legal search index, specialized instructions for legal analysis and writing, and integrations designed for law firms and legal technology companies.
The launch is aimed at workflows that go beyond general-purpose AI assistance, including legal research, contract analysis, deal diligence and drafting. OpenAI says API customers such as Harvey and Legora will be able to build Astra for Law into their own legal products and workflows.
What is OpenAI Astra for Law?
Astra for Law is GPT-6 Astra configured for professional legal work. It combines the underlying model with a dedicated Legal Search Index and custom instructions designed to help with legal research, analysis and writing.
The system is designed to help lawyers research questions using client facts, develop arguments and deal terms, and identify objections or weaknesses in an argument.
OpenAI’s configuration can also distinguish between a court’s actual holding and other observations in an opinion, identify cases that weaken an argument and explain how contractual exceptions can shift risk between parties.
Dedicated legal search covers U.S. legal materials
One of the main differences between Astra for Law and general web search is its dedicated legal search infrastructure.
OpenAI says the Legal Search Index covers more than 230 million URLs and is updated with new sources daily. The index includes:
- Case law
- Statutes
- Regulations
- Court rules
- Administrative decisions
The research experience also incorporates case-law material from the Free Law Project’s CourtListener collection, which OpenAI says covers more than 99.9% of published U.S. precedential case law.
OpenAI says the legal index is intended to complement licensed specialist products already used by law firms, including services from providers such as Thomson Reuters.
Astra for Law shows gains in legal research benchmark
OpenAI evaluated the complete Astra for Law configuration using 200 U.S. legal research questions from Vals AI’s Legal Research Bench.
At the highest reasoning setting, 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.
OpenAI describes that as a 40% relative improvement. The company also reported that Astra for Law found 24% more reference cases on case-law-focused questions and retrieved up to 54% more relevant passages from correct court opinions on an audited set of target passages.
These figures are OpenAI’s reported evaluation results, rather than an independent assessment of the system’s performance across legal practice.
Legal workflows extend beyond research
Astra for Law is designed to apply research to practical legal workflows.
According to OpenAI, its specialized instructions can help lawyers:
- Apply research to client-specific facts
- Develop legal arguments
- Analyze deal terms
- Identify weaknesses and uncertainty
- Review contractual exceptions
- Distinguish holdings from other judicial observations
- Find authorities that undermine an argument
OpenAI has also worked with selected law firms to develop customized interfaces and integrations around proprietary data and existing legal workflows.
Law firms are already building custom AI workflows
OpenAI highlighted several examples of firms developing legal applications around its technology.
Sullivan & Cromwell built an agreement analyzer using its negotiation playbooks and selected precedents. The system identifies risks that arise when contract provisions are considered together and produces proposed redlines and draft client advice for lawyers to review.
Ropes & Gray developed a deal-diligence system that works with data-room information and helps identify issues relevant to acquisitions, including contracts that could require notice or consent.
Cooley developed GO Public around its capital-markets expertise, covering work from IPO filing preparation through identifying risks requiring management attention.
The examples illustrate how OpenAI expects law firms to combine general AI capabilities with firm-specific knowledge, workflows and review procedures.
Astra for Law launches with 26 legal plugins
OpenAI is also launching 26 partner-built plugins connecting ChatGPT with specialist software used by legal professionals.
The integrations include tools such as:
- iManage, for working with matter files
- Intapp, for surfacing activities that may require time-entry review
- DeepJudge, for comparing prior deals
- Thomson Reuters, including HighQ matter context and a preview of a CoCounsel Legal connector
- Relativity
- Clio
The launch also includes nine community plugins created by lawyers and legal engineers from LegalQuants, LECG and Skills.law. Together, those community plugins provide 47 custom skills that practitioners can adapt for their workflows.
ChatGPT for Word is now generally available
Alongside Astra for Law, OpenAI says ChatGPT for Word is generally available for legal professionals.
The integration can help lawyers proofread documents, suggest edits and flag formatting issues directly within Microsoft Word.
This gives legal teams another way to incorporate AI into an existing document workflow rather than moving documents entirely into a separate application.
Privacy and governance features for law firms
Confidentiality and access controls are particularly important for legal organizations because law firms routinely work with sensitive client information.
OpenAI has created a Trusted Access Program for eligible law firms. The program provides access to Astra for Law for professional legal work and includes additional controls for participating firms.
For eligible firms, OpenAI says the offering includes:
- Zero Data Retention on the API
- ChatGPT Enterprise usage excluded from human review by default
- Work on information permissions and ethical walls
- Client instructions and firm oversight
OpenAI is also working with Latham & Watkins on governance approaches involving permissions, ethical walls and oversight.
Astra for Law availability
Astra for Law is initially being offered to selected law firms in the United States through OpenAI’s Trusted Access program, ChatGPT and Codex. OpenAI says the API version is coming soon.
In ChatGPT, the model appears as GPT-6 Astra Law, while the API model name is gpt-6-astra-law.
OpenAI says eligible lawyers and people working under their supervision can request early access through their OpenAI account team or OpenAI Sales.
How Astra for Law differs from regular GPT-6 Astra
The underlying GPT-6 Astra model is designed for broad professional work, including computer use, research, coding and document creation. Astra for Law adds a legal-specific layer around that model.
| Feature | GPT-6 Astra | Astra for Law |
| General reasoning | Yes | Yes |
| Computer use | Yes | Yes |
| Legal research | General-purpose | Dedicated legal search |
| U.S. case law | Web/search based | Legal Search Index + CourtListener |
| Legal drafting | General | Specialized legal instructions |
| Legal plugins | General ecosystem | 26 partner-built legal plugins |
| Law-firm workflows | Custom | Legal-specific configurations |
| Availability | Broad rollout | Initially selected U.S. law firms |
The key distinction is therefore not simply the underlying model. Astra for Law combines GPT-6 Astra with specialized legal data access, instructions, integrations and governance controls.
What Astra for Law means for legal AI
The launch shows OpenAI moving further toward domain-specific AI systems rather than relying solely on general-purpose models.
For law firms, the potential value lies in combining model reasoning with structured legal search, firm knowledge and existing software. At the same time, legal professionals remain responsible for reviewing research, sources and generated work before relying on it.
OpenAI’s own Help Center specifically advises users to review Astra for Law’s answers and cited sources before relying on them.
As the system expands to more firms and eventually the API, its integrations and specialized legal workflows could become an important part of how professional legal teams experiment with AI-assisted research and document work.
