OpenAI on Thursday announced Astra for Law, a new offering for legal work built on its most advanced and expensive model.
The product adds plugin tools, a way to search legal cases more effectively, and custom model tweaks designed to help others build on the tech. OpenAI said the tweaks ground the system in 'actual case law,' which it said led to fewer hallucinations and 'much more sound legal reasoning.'
The company is aiming the product at the legal tech industry and the 200 largest American law firms, known as AmLaw 200, according to Jason Boehmig, who leads legal product at OpenAI. Boehmig said the company is focused on 'the customer and not the competition.'
OpenAI said the tool gives answers that are about twice as long as typical Astra responses, and Boehmig said that was based on feedback that 'lawyers like longer responses.'
The launch comes as OpenAI prepares for its initial public offering and keeps pushing into industry-specific products. The company last week unveiled ChatGPT for Financial Services for Wall Street customers. It has also launched products for teachers, clinicians, life sciences researchers, and the personal health and finance sectors.
OpenAI’s move lines up with a legal release from Anthropic in May, though OpenAI said its own product is tailored specifically for legal work by embedding it with tools and tweaking its settings.
The launch also includes plugins for ChatGPT from legal tech companies Harvey and Legora, plus Thomson Reuters, iManage, DeepJudge, and others. OpenAI said it will bring a citations feature from its financial services launch to the legal offering soon.
For founders, the signal is clear: the AI race is moving from general-purpose chat to vertical products that can charge for deeper workflow integration. That is where pricing power lives. If OpenAI can make legal output more relevant and less error-prone, it is not just selling software — it is selling trust inside one of the most expensive knowledge-work markets.
That matters because enterprise customers do not buy novelty for long. They buy tools that save time, cut friction, and fit into existing workflows. Founders who get this right move faster, and they keep more of the margin when the product becomes part of the daily workflow.
