Anthropic AI Updates: July 24, 2026

1. Anthropic Expands Its Claude Certification Program to Four Role-Based Credentials

Anthropic. Anthropic expanded its Claude Certification Program to four role-based credentials, three of them new: Claude Certified Associate Foundations for daily business and technical users, Claude Certified Developer Foundations for engineers building on the API and agents, Claude Certified Architect Foundations for designing Claude-based agent systems, and an advanced Claude Certified Architect Professional covering integration, governance, and evaluation at enterprise scale. The exams are proctored through Pearson Professional Assessments with identity verification, and the company said more than 36,000 consultants have been certified across 1,300 organizations since March, with firms including Accenture, PwC, and Deloitte committing to certify tens of thousands of employees. Certified practitioner counts now factor into a partner’s tier standing in the Claude Partner Network. Source

2. Claude Voice Mode Now Runs on Opus and Sonnet With Tool Actions

Anthropic. Anthropic extended Claude’s voice mode so users can work through longer problems out loud, running the feature on Claude Opus, Sonnet, and Haiku rather than Haiku alone and letting people switch models mid-conversation for deeper reasoning. Voice can now trigger actions in connected services such as Gmail, Slack, Google Calendar, and Canva, for example rescheduling a calendar event after talking through a decision. The update adds support for 11 languages and is available on mobile, desktop, and web, with free plans limited to Haiku and a single tool. Source