Anthropic debuts Claude Sonnet 5.5 running 30% faster than the previous-generation AI model

Anthropic unveiled Claude Sonnet 5.5, a mid-tier model in their AI lineup, boasting significant improvements over its predecessor. This new model is geared towards everyday tasks, offering a faster, more cost-effective solution. Operated as the workhorse for Anthropic's Claude family, Sonnet 5.5 excels in document generation, summarization, spreadsheets, and coding bug fixes.

The model possesses a keen design sense, capable of enhancing user interfaces and creating polished slides with minimal editing. It also features clearer writing styles and human-like communication, making it an ideal partner for collaborative tasks.

In comparison to the previous generation, Sonnet 5.5 runs 30% faster and costs less, priced at the same rate as its predecessor. This model integrates a text watermarking system to detect AI-generated text, ensuring compliance with global regulations like the European Union's AI Act. Early testers noted that Sonnet 5.5 is faster in iterations for less complex tasks and more intelligent than its predecessor.

Companies like Zendesk have found Sonnet 5.5 to be highly efficient, reducing ticket processing time by 20% and making fewer wrong decisions. In coding evaluations, Sonnet 5.5 scored 70.6% on Terminal-Bench 4.0, outperforming the previous model's 10.3%. This model is the first Sonnet to surpass Pokémon Red using only screenshots, showcasing its capabilities in long-horizon work and image understanding.

Sonnet 5.5 is the fastest Sonnet model to date, with output speed enhancements. The company plans to release Haiku 5.5, the smallest, fastest, and most efficient model in the Claude family, in the coming weeks. This model, built on the same foundation, will also feature notable cybersecurity capabilities, making it the first Sonnet model with cybersecurity safeguards and fallbacks for sensitive content.