ZeroDrift prices AI compliance at $0.01 a message

ZeroDrift launched Anchor 3.0, claiming 95.5% recall on 150 human-labeled FINRA compliance tasks at $0.01 per message. ZeroDrift’s Anchor 3.0 model page says the system runs up to 34 times faster and 12 times cheaper than four frontier models, while marking those results self-reported and noting an independent run is in progress. Regulated-AI teams can price a message-level enforcement layer before granting agents send access; Palo Alto’s $1T security-debt thesis is a reminder to measure controls rather than trust vendor market sizing.