On June 21, 2024, Anthropic released Claude 3.5 Sonnet, the first model in its planned Claude 3.5 family. Anthropic said the model “raises the industry bar for intelligence, outperforming competitor models and Claude 3 Opus on a wide range of evaluations,” while operating “at twice the speed of Claude 3 Opus.” It kept the same $3-per-million-input-token, $15-per-million-output-token pricing and 200K-token context window as the earlier mid-tier Sonnet.
The launch emphasized coding and vision. Anthropic reported that Claude 3.5 Sonnet solved 64 percent of problems on an internal agentic coding evaluation, compared with 38 percent for Claude 3 Opus, and called it “our strongest vision model yet, surpassing Claude 3 Opus on standard vision benchmarks.” Alongside the model, Anthropic introduced Artifacts, a feature where generated content like code, documents, or website designs “appear in a dedicated window alongside their conversation,” turning Claude from a chat box into a shared workspace.
Claude 3.5 Sonnet mattered because it inverted the usual scaling assumption: a smaller, cheaper, faster model outscored the previous flagship. That made strong AI dramatically cheaper to deploy at scale and set the pattern, repeated across the industry, of mid-tier models carrying most real production traffic.