Anthropic released Sonnet 5.5 on September 28, promising faster iteration when you're building an app or fixing bugs. It reports output generation over 30% faster than Sonnet 5.
In early testing across 118 app builds, Base44 reported that Sonnet 5.5 matched Opus 5's quality while averaging 3.6 iterations per build, compared with 7.7 for Opus 5. Those results come from a launch partner featured in Anthropic's announcement.
API prices stay at $2 per million input tokens and $10 per million output tokens. Anthropic reports up to 30% lower task costs because Sonnet uses fewer tokens. That doesn't mean your Claude subscription becomes 30% cheaper.
Sonnet 5.5 is available in Claude apps and Claude Code. Anthropic positions it for clearly defined everyday work, including documents and slides. It still recommends Opus 5.5 for complex, open-ended tasks requiring sustained judgment.
For existing apps using the API, switching models may require configuration changes. The migration guide covers compatibility checks before the switch.