Anthropic releases 20+ legal AI integrations and Claude Opus 4.7 model scoring 90.9% on BigLaw benchmark, as major law firms adopt the tools despite ongoing hallucination risks

Anthropic announced more than 20 new integrations with legal software, including 12 role-specific AI plugins for tasks like M&A due diligence and employment handbook drafting, plus a cross-app integration with Microsoft 365 embedding Claude across Word, Outlook, Excel, and PowerPoint. Law firms Freshfields, Quinn Emanuel Urquhart & Sullivan, Holland & Knight, and Crosby Legal are using Claude on live matters.

The underlying model, Claude Opus 4.7, scored 90.9% on Harvey's BigLaw Bench, the legal industry's benchmark for AI legal test uses. Anthropic's approach to reducing hallucinations relies on "grounding"—constraining Claude to draw only from live, verified sources like Westlaw case law databases and CourtListener's court opinion archives, rather than generating answers from training data memory.

Legal AI products including Harvey, Legora, Solve Intelligence, and Eve are built on Claude's underlying models. More than 20,000 legal professionals signed up for Anthropic's most recent legal webinar, described by the company as the largest legal session it has ever held.

Anthropic is partnering with Free Law Project, Courtroom5, and other legal aid organizations to provide access to Claude connectors at no additional cost, positioning the expansion as addressing access to justice alongside serving BigLaw; roughly 80% of civil litigants appear in court without a lawyer, according to Anthropic.