Claude Sonnet 5.5

Claude Sonnet 5.5, the second model in the Claude 5.5 family, has been introduced as a faster, lower-cost alternative to Claude Sonnet 5. This new model excels at well-defined everyday tasks, such as bug fixing and creating polished documents, slides, and spreadsheets. It also boasts a strong design sense. Sonnet 5.5 outperforms its predecessor by 30% in speed and up to 30% in cost for most work.

The model scores 70.6% on Terminal-Bench 4.0, an agentic coding evaluation, compared to Sonnet 5's 10.3%. It also scores two points below Opus 5.5 on GDPval-AA, a test of real-world work across various occupations. However, Sonnet 5.5 surpasses Pokémon Red working only from screenshots, making it the first Sonnet model to excel at long-horizon work and image understanding.

In terms of collaboration, Sonnet 5.5 is described as a better partner than Sonnet 5, writing more clearly. Its speed makes it suitable for fast iteration on less complex tasks. Sonnet 5.5 is priced at the same rate as Sonnet 5, but typically requires fewer tokens, resulting in up to 30% lower costs per task. It is also faster, generating outputs 30%+ quicker than Sonnet 5.

On automated behavioral audits, Sonnet 5.5 matches or improves upon Sonnet 5 in most alignment measures. It includes cybersecurity capabilities comparable to Opus 5, making it the first Sonnet model with cyber safeguards. Sonnet 5.5 outperforms Sonnet 5 in several evaluations, often performing comparably to Opus 5.5. However, Opus 5.5 remains more robust at complex, open-ended work requiring sustained judgment.