Claude 3.5 Sonnet

Launched June 20, 2024, Claude 3.5 Sonnet was Anthropic's mid-tier model that outperformed the larger Claude 3 Opus on most benchmarks while running at twice the speed — then became 2024's default coding model. In Anthropic's internal agentic coding evaluation it solved 64% of problems versus Opus's 38%, and it kept a 200K context window at $3/$15 per million input/output tokens.

The October 2024 update pushed SWE-bench Verified from 33.4% to 49.0% and introduced computer use in public beta — letting the model view screenshots and operate a mouse and keyboard — making it one of the first widely available models built for agentic computer operation. It remained Anthropic's workhorse through early 2025 before the Claude 3.7 and Claude 4 lines. Proprietary, available via API, Claude.ai, AWS Bedrock, and Vertex AI.