Claude 3.5 Sonnet raises the coding bar

Claude 3.5 Sonnet outperforms Claude 3 Opus at a fifth of the cost and solves 64% of problems in an internal agentic coding eval.

What happened and why it matters

Claude 3.5 Sonnet became the default model for coding agents, more than doubling Claude 3 Opus on Anthropic's agentic coding evaluation and kick-starting the agentic-coding tooling boom.

Sources

• Introducing Claude 3.5 Sonnet ↗ Anthropic · 20 Jun 2024