Anthropic says Claude AI designed protein binders for 14 of 15 targets in lab

Anthropic says Claude AI designed protein binders for 14 of 15 targets in lab

tests

The company reports its AI models autonomously designed drug-candidate proteins at hit rates roughly double the industry norm.

Anthropic said its Claude AI models produced 354 confirmed protein binders across 14 of 15 targets in wet-lab testing by Adaptyv Bio and Twist Bioscience. anthropic

Hit rates ranged from 22% to 35% depending on setup, roughly double the 10-15% typical in human-led protein design campaigns, according to the company. anthropic

Anthropic acknowledged the designs are not drugs and flagged dual-use safety risks, saying advanced life-science capabilities remain restricted in its most powerful model