OpenAI Launches Astra for Law, Bringing GPT-6 Astra to Legal Research with 230 Million Case Law URLs

5 Sources

Share

OpenAI unveiled Astra for Law, a legal-focused AI platform built on GPT-6 Astra that helps law firms conduct research and draft legal advice. The platform indexes over 230 million URLs of U.S. case law and achieved 54% accuracy on legal research benchmarks. But concerns remain about AI hallucinations in legal work.

News article

OpenAI Enters Legal AI Market with GPT-6 Astra Platform

OpenAI released Astra for Law on Thursday, marking its aggressive entry into the competitive legal AI market

1

. The legal-focused AI platform combines the company's GPT-6 Astra model with specialized tools designed for law firm users and legal software providers

2

. Astra for Law is not an entirely new model but rather a configuration of GPT-6 Astra wrapped in a comprehensive legal search index covering more than 230 million URLs of U.S. case law, statutes, regulations, court rules and administrative decisions

5

. The platform draws its case law from the nonprofit Free Law Project's CourtListener database, which encompasses more than 99.9% of published U.S. precedential case law

2

.

Performance Benchmarks Show Significant Accuracy Gains

Astra for Law demonstrated substantial improvements in legal research capabilities during testing. The platform passed overall correctness checks on 54% of 200 questions drawn from the private validation set of Vals AI's Legal Research Bench

2

. By comparison, GPT-6 Astra with web search alone managed only 38.7% on the same questions when both systems ran at their highest reasoning effort

5

. On case-law-focused questions specifically, Astra for Law found 24% more reference cases than its predecessor

2

. The platform retrieved up to 54% more relevant passages from correct court opinions on a separate audited set, demonstrating its enhanced ability to locate precise legal authority

5

.

Major Law Firms and Legal Software Integrations

OpenAI partnered with prominent law firms including Sullivan & Cromwell, Ropes & Gray, Cooley, Latham & Watkins and Wachtell Lipton during the testing and development phase

1

. Sullivan & Cromwell developed an agreement analyzer that incorporates negotiating playbooks and precedents into deal reviews, generating proposed redlines

2

. Ropes & Gray focused on deal diligence tools, while Cooley built GO Public for initial public offering preparation and filing drafts

2

. Legal AI vendors including Harvey and Legora will build products on Astra for Law through API access

3

. The platform includes integrations with legal software providers such as Relativity, Clio, Intapp, iManage and Thomson Reuters

1

. Twenty-six partner-built plugins launched alongside the model, with Thomson Reuters bringing HighQ matter context into ChatGPT and previewing a CoCounsel Legal connector

2

.

Privacy Protections and Controlled Access for Confidential Work

OpenAI structured initial access through a Trusted Access program in ChatGPT and Codex, where selected law firms can access the platform with protections for confidential client work

1

. For eligible firms, the offering includes Zero Data Retention on the API, and ChatGPT Enterprise usage is excluded from human review by default

2

. OpenAI is collaborating with Latham & Watkins on governance design for information permissions, ethical walls and client instructions

5

. An API version called gpt-6-astra-law is scheduled for release later, though no specific date or pricing has been announced

2

.

Intensifying Competition in Legal AI Space

The launch of Astra for Law escalates competition among technology companies racing to capture business from law firms and legal departments

1

. Google expanded its Gemini Enterprise offerings for legal professionals last month, while Anthropic has released tools for lawyers using its Claude AI assistant since January

3

. Thomson Reuters launched an AI model called Thomson 1.0 last month, trained on its legal research content and built on open-source technology drawing from decades of authoritative legal content

1

. Joel Hron, chief technology officer at Thomson Reuters, emphasized that legal professionals need more than connectivity, requiring trusted intelligence, relevant enterprise and matter context, purpose-built legal capabilities, and governance for high-stakes work

2

.

Concerns About AI Accuracy and Hallucinations Persist

Despite technological advances, concerns about AI for lawyers remain prominent in the legal community. Chief Justice John Roberts previously cautioned that AI-generated predictions could influence judges and potentially challenge judicial independence

4

. George Clooney revealed at a Venice Film Festival symposium that his wife, human rights lawyer Amal Clooney, has encountered AI-fabricated laws, warning that AI can "hallucinate and make up fake cases" that lawyers or judges may cite without verification

4

. These hallucinations pose significant risks in legal analysis where accuracy is paramount. Watch for how OpenAI addresses these accuracy concerns as adoption grows, whether competing platforms can match Astra for Law's benchmark performance, and how regulatory frameworks evolve to govern AI use in legal practice. The platform's long-term success will depend on maintaining trust while demonstrating consistent reliability in high-stakes legal work where errors carry serious consequences for clients and practitioners alike.

Today's Top Stories

© 2026 TheOutpost.AI All rights reserved