Anthropic Releases Sonnet 5.5: Faster and Cheaper Than Its Predecessor

Anthropic has unveiled Sonnet 5.5, a new version of its mid-tier model. The company frames it as a handy assistant for everyday tasks, from writing code to producing office documents.

The previous release, Sonnet 5, arrived about three months ago, and its main draw was efficient agentic deployment — the ability to run agents at a lower cost than competitors. For Sonnet 5.5, the headline argument is speed. Anthropic claims the model is 30% faster than its predecessor, with a significantly slower rate of token burn.

Where it sits in the lineup

In Anthropic's hierarchy, Sonnet is less powerful than the Opus model, but in certain situations it proves more useful thanks to its agility. The company's own benchmarks show Sonnet 5.5 outperforming Opus 5.5 on agentic coding — likely because it can spawn multiple agents without breaching cost limits.

Cyber capabilities and safeguards

The new model is also credited with notable cyber capabilities. Anthropic says they are "comparable" to those of Opus 5. As a result, Sonnet 5.5 becomes the first Sonnet model subject to the same cyber safeguards that already apply to Fable and Opus.

The company also plans to refresh Haiku, its smallest model, in the coming weeks, though it gave no firm date.

The race goes on

The past year has brought a flood of announcements from the major labs. Just last week OpenAI released several new models, including enhanced versions of Sol and Luna, its mid-tier and budget offerings. Meta, meanwhile, announced a model meant to power an upcoming feature for its smart glasses.

Source: TechCrunch