.svg){kind=icon}
What is Claude Sonnet 5?
Claude Sonnet 5 is a mid-tier AI model from Anthropic, released on June 30, 2026. It sits between the lightweight Haiku and the flagship Opus in Anthropic's lineup — the balanced option, more capable than Haiku but priced well below Opus. Despite that, Sonnet 5 performs close to Opus 4.8, both in benchmarks and in real-world tasks, which is why Anthropic made it the default model on free and paid plans.
Founded in 2021 by former OpenAI researchers Dario and Daniela Amodei, Anthropic is the company behind Claude. Many consider it the AI company to watch. Enterprise clients love Anthropic's models — and for good reason. The company creates highly reliable models that produce top-tier results. They are especially well-regarded as the best AI coding models.
Sonnet 5 pushes the price-to-performance ratio further. It launches at $2 per million input tokens and $10 per million output tokens through August 31, then $3/$15 — undercutting flagship models that have grown slower and pricier with each release. Opus 4.8, by comparison, runs $5 per million input and $25 per million output.
Claude Sonnet 5 Key Upgrades
Sonnet 5 keeps its 1-million-token context window and adds adaptive thinking with selectable reasoning effort — low, medium, high, max, and extra-high — so it spends more effort only when a task needs it.
It's the most agentic Sonnet yet: stronger planning, tool use, and autonomous execution, scoring 57.4% on Humanity's Last Exam (HLE) and 81.2% on OSWorld-Verified for computer useOpenClaw.
Like all Claude models, Sonnet 5 is proprietary. It can't be run locally or self-hosted. One of the easiest ways to try Sonnet 5 is through Overchat AI, which doesn't require a separate Anthropic account or API key.
Anthropic's models are known for being direct about their capabilities and limits, and Sonnet 5 lowers hallucination and sycophancy rates further than 4.6. It will push back on incorrect information and may even correct you — which makes for more productive sessions.
