Anthropic AI Updates: June 7, 2026

1. Garvan Institute Uses Claude Across Genomics, Diagnosis, and Drug Discovery

Anthropic. Anthropic published a customer case study describing how Australia’s Garvan Institute of Medical Research uses Claude across research, operations, and administration, with more than 20 engineers and data scientists working with agentic development tools. The institute applies Claude to analyze hundreds of terabytes of genomic data, build multi-agent systems for genetic variant interpretation that compress rare-disease diagnosis from years toward hours, and support drug discovery work. The study reports that researchers have identified mechanisms explaining roughly 40% of disease genetic factors at the cellular level, with individual researchers able to interrogate knowledge across fields by managing AI agents rather than coding manually. Source

2. delight.ai Builds Debugging and Regression Tooling With Claude Code

Anthropic. Anthropic published a case study on delight.ai, which builds customer-support AI agents on Sendbird’s messaging infrastructure and handles billions of conversations monthly. The engineering team uses Claude Code to build internal tooling, including a unified debugging tool that fetches conversation logs and surfaces issues and a regression-testing system that generates test conversations at scale before deployment. The company reports production fixes dropping from about a week to one or two days and weekly pull requests roughly doubling since adopting the tooling in November, running a multi-model setup on Amazon Bedrock and the Anthropic API with Claude Opus for high-accuracy analysis. Source