Jun 2024
Claude 3.5 Sonnet — New State of the Art at Mid-Tier Cost
Anthropic released Claude 3.5 Sonnet, which outperformed Claude 3 Opus across most benchmarks while running at Sonnet-tier speed and pricing — a significant efficiency gain. It achieved 49% on the SWE-bench Verified coding evaluation, surpassing all other models at the time. Anthropic simultaneously introduced Artifacts, an interactive canvas panel within Claude.ai for viewing, editing, and running generated code and documents in real time.
Why it mattersReaching Opus-level quality at Sonnet pricing meant the quality-vs-cost tradeoff shifted: most production workloads no longer needed to compromise on capability.
Try itRe-benchmark your current model choice against Claude 3.5 Sonnet on your actual production prompts — many teams that stayed on GPT-4 found 3.5 Sonnet faster and cheaper for the same quality.