Claude Opus 5

by Claude model

Anthropic's current model for, in its own words, "complex agentic coding and enterprise work", and the one its docs tell you to start with. 1M-token context window, 128k max output, adaptive thinking, and a May 2026 training cutoff — the most recent of any Claude model. $5 per million input tokens and $25 per million output. Claude Fable 5 sits above it for workloads needing the highest available capability, at double the price.

Visit site

Good at