Claude Sonnet 5

by Claude model

Anthropic's mid-tier model, positioned in its docs as "the best combination of speed and intelligence" — the default choice when Opus is more capability than the task needs. 1M-token context window, 128k max output, adaptive thinking, and $2 per million input tokens against $10 per million output, which is under half what Opus 5 costs. January 2026 training cutoff.

Visit site

Good at