What is Sonnet?

Sonnet is Anthropic’s most widely used AI model. It sits in the middle of their model lineup: Haiku (fastest, cheapest), Sonnet (balanced), and Opus (most capable). Sonnet gives you strong performance without the cost or latency of Opus, which is why it’s the default choice for most people.

Anthropic names their models after literary forms. Sonnet is the middle tier, substantial and well-rounded. Each tier has different versions as the models get updated. As of February 2026, the latest is Sonnet 4.6, which came out on February 17, 2026 and is priced the same as Sonnet 4.5.

Sonnet pricing is significantly cheaper than Opus. For Sonnet 4.5, it’s around $3 per million input tokens and $15 per million output tokens (exact pricing varies). That makes it affordable enough to use all day while still being highly capable for most tasks like coding, writing, analysis, and conversation.

What makes Sonnet the go-to choice is the balance. It’s fast enough that you don’t feel like you’re waiting. It’s smart enough to handle complex coding tasks, reason through problems, and generate quality output. And it’s cheap enough that you can use it freely without worrying about costs. For 95% of what people do with AI, Sonnet is plenty.

The tradeoff is that when you hit a really difficult task, Opus will do better. But for everyday work, Sonnet is the sweet spot.

The latest features I shipped for Code Q&A were built using Sonnet through Claude Code. I get to use Opus via Claude Code at my day job.