tech
Anthropic launches Claude Sonnet 5, a cheaper agent model
Anthropic has launched Claude Sonnet 5, its most agentic mid-tier model yet. It runs close to the flagship Opus 4.8 on many tasks, but costs less than half as much.
TL;DR
- Anthropic launched Claude Sonnet 5, its most agentic mid-tier model.
- Sonnet 5 offers performance close to the flagship Opus 4.8 at less than half the cost.
- The model is built to act, make plans, drive browsers and terminals, and run autonomously.
- Its pricing starts at $2/million input tokens and $10/million output tokens (introductory rate until Aug 31, 2026).
- The high cost of running AI agents is a key problem Sonnet 5 aims to solve.
- A new tokenizer means text can map to more tokens, requiring careful cost monitoring.
- On agentic coding tests, Sonnet 5 scored 63.2%, compared to Opus 4.8's 69.2%.
- Sonnet 5 shows improved safety, hallucinating and flattering less than its predecessor.
- It has real-time cyber safeguards enabled by default, similar to Opus models.
- The model is available across Anthropic's apps, Claude Code, and the API.