Claude Sonnet 5
Claude Sonnet 5 is Anthropic’s most agentic Sonnet model yet, delivering performance that rivals Opus 4.8 while costing less. Benchmarks show it narrows the gap on reasoning, tool use, coding and knowledge work, achieving scores near Opus 4.8 on the BrowseComp and OSWorld‑Verified evaluations. Safety tests show fewer undesirable behaviors than Sonnet 4.6, though cybersecurity skill stays weaker than Opus. The model runs autonomously, can plan, use browsers and terminals, and self‑correct without explicit prompting.
The model is now the default for Free and Pro tiers and is available to Max, Team and Enterprise users via the Claude API as claude-sonnet-5. Introductory pricing runs through Aug 31 2026 at $2 per million input tokens and $10 per million output tokens, rising to $3 and $15 thereafter. Early partners say it finishes complex end‑to‑end tasks—updating Salesforce tiers, drafting launch announcements—without stopping short, and routinely checks its own output. One engineer noted it “gives our agents a strong execution layer for multi‑step software engineering work,” highlighting its ability to sustain coding, tool use and debugging at an attractive price.