Anthropic Deploys Sonnet 5 Model with Enhanced Agentic Capabilities and Security Protocols

The release of the Sonnet 5 AI model introduces expanded autonomous task execution, including integrated web browsing, multi-step planning, and direct terminal access for developers. This iteration aims to bridge the performance gap between mid-tier and flagship models while maintaining lower operational costs for enterprise and individual users.

Security architecture has been updated to improve resilience against prompt-injection attacks and unauthorized manipulation. The model incorporates refined safety filters designed to detect and block cybersecurity-related misuse, replacing previous, more restrictive protocols.

Deployment is currently active across all user tiers, including free, professional, and enterprise accounts. API pricing is structured with an introductory rate effective through August 2026, after which costs will increase by 50 percent for both input and output tokens.