Anthropic has refreshed the middle of its Claude lineup, replacing Sonnet 5, which arrived in June, with a new Sonnet 5.5 model. The update comes just a week after Opus 5.5 landed, when the company said upgraded Sonnet and Haiku models would follow within weeks. Haiku is now the only one of the three core models still waiting for its turn.
For anyone unfamiliar with how Claude is structured, the family scales by size and cost. Haiku is the smallest and cheapest, Sonnet sits in the middle, and Opus is the largest and priciest of the three. Above all of them is the Fable and Mythos tier, which Anthropic positions as its most capable and most expensive option. Sonnet has typically been the model many developers and businesses settle on, since it balances capability with a cost that makes sense at scale.
Anthropic is pitching Sonnet 5.5 at well-defined, everyday work rather than sprawling open-ended projects. The company says it performs best on tasks such as fixing bugs and producing finished documents, slides and spreadsheets. It’s also leaning on design work, claiming the model can refine user interfaces and build slides that follow existing templates with little cleanup needed afterwards. According to Anthropic, early testers found it a better collaborator than Sonnet 5, with clearer writing and enough speed for rapid back-and-forth on simpler jobs. Those are the company’s own characterisations, and independent testing will show how well they hold up.
Pricing hasn’t changed. Sonnet 5.5 costs $2 per million input tokens, $10 per million output tokens and $0.20 per million cache reads, matching its predecessor. The more interesting claim is efficiency. Anthropic says the new model usually needs far fewer tokens to finish the same task and generates output more than 30 percent faster than Sonnet 5, which would make it the quickest Sonnet yet. If the token savings are real, identical pricing effectively becomes a price cut for many workloads.
There’s also a notable shift on safety. Anthropic says Sonnet 5.5 has cybersecurity capabilities comparable to Opus 5, and as a result it’s the first Sonnet model to ship with cyber safeguards and fallbacks similar to those applied to the company’s top-tier models. Biology safeguards stay the same as Sonnet 5. Anthropic says both sets of protections are aimed at a narrow range of high-risk requests, and that routine software development and most life sciences work shouldn’t be affected.
That detail says a lot about where the industry is heading. When a mid-tier model reaches the security capabilities of last generation’s flagship, the protections once reserved for the top end start moving down the lineup too.
