Meet Claude Sonnet 5: The Most Agentic Sonnet Model Yet

Anthropic has released Claude Sonnet 5, a next-generation model built to be the most agentic Sonnet yet. It can make plans, use tools like browsers and terminals, and run autonomously at a level that previously required larger, more expensive models.

We are thrilled to announce that Claude Sonnet 5 is now live on VEGA AI. Here is a breakdown of why this update is a major milestone for developers and professionals alike.

Narrowing the Gap to Opus

For a long time, the clearest gains in autonomous execution were reserved for the flagship Opus models. Sonnet 5 narrows that gap: its performance is close to that of Opus 4.8, but at lower Sonnet pricing. It represents a strict improvement over Sonnet 4.6 on important agentic benchmarks like reasoning, tool use, coding, and knowledge work.

Key Highlights & Benchmark Gains

Early testers consistently report that Sonnet 5 is much more autonomous and resilient:

Sustained Coding: Sonnet 5 handles complex programming tasks, writes tests, and refactors dirty legacy code with ease. It can trace a failure to its actual root cause and ship a durable fix in a single pass.

Multi-Step Tool Use: Whether navigating a browser or writing terminal commands, Sonnet 5 stays on plan and executes complex sequences of actions with high precision.

Self-Correction: The model checks its own outputs unprompted, identifying and correcting errors before delivering the final result.

Enhanced Safety

Safety assessments show that Sonnet 5 has a lower rate of undesirable behaviors than Sonnet 4.6, making it safer to deploy in autonomous agent contexts. It cleanly and consistently refuses unsafe requests while maintaining high helpfulness for legitimate tasks.