Faster, safer, smarter, cheaper: Anthropic makes the case for Claude Sonnet 5.5

With as fast as things move in AI, you can’t just innovate along one narrow axis and expect to get ahead. Outshining your competition means delivering palpable gains across a whole spectrum of angles — and somehow, the biggest AI models around keep managing to do that, nearly week after week. We already saw how Claude Sonnet 5 delivered some real benefits over Opus 4.8, and now Anthropic is back to hype up Claude Sonnet 5.5.

It’s been just over a week now since Anthropic introduced Claude Opus 5.5, hitting some new performance and communication benchmarks in the process. Sonnet 5.5 is basically the economy version of that, prioritizing speed and offering a significantly lower cost of operation — without sacrificing too much on power.

Looking directly at Sonnet 5, the new 5.5 release should be at least 30% faster, while keeping its same lower-than-Opus pricing and beefing up security. It also hits new highs for the model across performance benchmarks — including the feather in its cap of being able to complete Pokémon Red.

While Sonnet 5.5 won’t beat Opus 5.5 when it comes to the heavy lifting, it’s definitely holding its own a lot better than you might expect:

Beyond just trying to hit hard on those “faster, cheaper, better” deliverables, Anthropic also talks about the steps it’s taking to not just harden Sonnet against attempts to use it for harm, but also to steal it outright — this is the first time the model has been empowered to resist efforts to extract it through distillation attacks, helping to keep Anthropic’s secret sauce safe.

Claude Sonnet 5.5 is now generally available across platforms providing access to Anthropic models.