September 18, 2026, (Inside AI) — OpenAI introduced a legal industry version of its GPT-6 Astra model on Thursday, September 17, 2026, marking its most aggressive push into professional services. The Sam Altman-led company said the model includes settings, tools, and context tailored for legal work, initially available to a handful of law firms through Trusted Access in ChatGPT and Codex.
The launch positions OpenAI against Anthropic, which expanded Claude's legal sector reach earlier this year with 12 practice-specific plug-ins. Both companies are racing to embed AI in legal workflows, where massive document volumes and repetitive tasks create strong demand. The global legal services market exceeds $900 billion annually, according to industry estimates, making it a prime target for AI automation.
Law Firms Gain Custom AI, But Judgment Stays Human #
OpenAI designed the new system to work alongside a firm's existing expertise and workflows rather than replace them. The company emphasized data protection and professional obligations as core features. Its legal products include zero retention for API usage, while ChatGPT Enterprise offers an 'eyes-off' human review posture by default. This addresses client confidentiality and independent judgment requirements that govern legal practice.
The tool connects OpenAI's models with software law firms already use. Partnerships with legal tech providers including Clio, DeepJudge, iManage, Intapp, Relativity, and Thomson Reuters allow firm knowledge, case information, documents, and established workflows to flow into ChatGPT while retaining existing permissions and controls. According to OpenAI, Astra for legal can support tasks from researching matters and analyzing documents to drafting and reviewing deals. Its legal industry page shows examples involving IPO work, document review, deal diligence, and the use of a firm's own precedents and playbooks.
Read: Orrick Hires CTO and CFO to Lead AI-Driven Legal Operations
Legal professionals have increasingly adopted AI for research, drafting, and document-heavy work. A 2025 survey by the American Bar Association found that 45% of law firms reported using AI tools, up from 30% the previous year. This growth has prompted bar associations and courts to issue guidance on AI use, with some requiring disclosure when AI assists in legal filings. OpenAI's approach of integrating with existing firm systems and respecting permissions may help law firms meet these evolving standards.
OpenAI's move also reflects a broader trend of AI companies targeting professional services. By combining AI reasoning with firm-specific knowledge, specialist databases, and professional tools, OpenAI aims to make its models part of established legal workflows. The company retains lawyers wherever legal judgment is required, signaling a collaborative rather than replacement strategy.
Anthropic's earlier expansion allowed Claude users to access 12 legal practice plug-ins covering litigation, employment, and commercial law, deployable through its Cowork platform or embedded into law firms' systems. That move similarly emphasized integration and customization. Both companies are betting that AI can augment legal work without displacing the nuanced judgment that defines the profession.
Inside AI could not independently verify the specific capabilities or performance of GPT-6 Astra for legal tasks. OpenAI did not disclose pricing or the exact number of firms in the initial rollout. The company also did not provide a timeline for broader availability. As AI continues to penetrate legal services, questions about liability, accuracy, and ethical use will likely shape adoption. For now, OpenAI's latest offering signals that the competition for legal AI is heating up, with law firms gaining powerful new tools and the responsibility to use them wisely.