OpenAI has launched a legal-focused version of its new GPT-6 Astra model that harnesses a US legal research database, drawing on work with firms including Sullivan & Cromwell (S&C), Latham & Watkins, Ropes & Gray and Cooley.
Astra for Law sets out to provide a “new foundation” which law firms and legaltech companies can use to develop their own AI tools and workflows.
It combines GPT-6 Astra with a legal research index covering US case law, statutes, regulations, court rules and administrative decisions.
It will initially be offered to selected law firms with API access to follow, while legal AI platforms including Harvey and Legora will be able to use the API to incorporate Astra for Law into their own products and workflows.
OpenAI said its engineers had teamed up with S&C to build an “agreement analyzer that brings the firm’s negotiating playbooks and selected precedents into the review of a new deal”.
Ropes & Gray has built a deal diligence system “around how its lawyers work through a data room”, while Cooley has built a tool “to bring its capital markets expertise into how companies prepare to go public”, OpenAI said.
S&C partner John Savva said: “Across both litigation and transactional matters, the models demonstrated impressive research depth and sensitivity to authority. Even at this early stage, they felt like a significant step toward legal-focused AI that is carefully grounded in research that is both current and comprehensive.”
Niko Grupen, head of applied research at Harvey, commented: “In our early testing, Astra for Law showed strength across key aspects of legal research: grounding answers in on-point authorities, citing with precision and offering practical, advisory guidance.”
The legal research index covers more than 230 million URLs, with sources added daily. It incorporates Free Law Project’s case-law collection, which covers more than 99.9% of published US precedential case law.
OpenAI has also introduced 26 partner-built plugins from providers including Thomson Reuters, Intapp, Harvey, Legora, DeepJudge and iManage, connecting ChatGPT with specialist software already used by law firms.
For example, with iManage a lawyer can draft a negotiation brief in ChatGPT and save it to a matter file, while Legora enables lawyers working in ChatGPT to access its legal research capabilities.
The launch also includes nine community-built plugins developed by lawyers and legal engineers, containing 47 custom skills that can be adapted for different legal tasks.
Latham & Watkins, meanwhile, is working with OpenAI on governance, evaluation and controls for advanced AI in legal work, including how firms can deploy increasingly capable models while maintaining professional standards.
The launch of Astra for Law follows OpenAI's hire in June of Jason Boehmig, the co-founder of AI contracting platform Ironclad, as product leader for its new legal vertical.
OpenAI’s plugins offer a similar toolkit to one launched by tech rival Anthropic in May this year, when it unveiled 12 practice area plugins via its Claude AI-agent technology. Those practice areas include commercial, corporate, employment and privacy law, among others.
Google, meanwhile, launched Gemini Enterprise for Legal in preview last month. It combines legal-specific AI tools and agents with connectors to law firms’ existing technology and data.
Brian W Tang, founding executive director of the University of Hong Kong’s LITE Lab, wrote on LinkedIn: “What a change in AI in legaltech in less than 3 months!”
He said a new “enterprise-grade LLMs targeting legal” category had effectively been created, pointing to Anthropic’s legal plugins for Claude, Google’s Gemini Enterprise for Legal and OpenAI’s Astra for Law. He asked whether this trend would add "to the already mounting Legal AI enterprise subscription costs of law firms and legal departments".
Last month, Thomson Reuters launched its first proprietary large language model, allowing it to move away from the high running costs of typical frontier models.
Email your news and story ideas to: [email protected]
