Anthropic Releases Claude Sonnet 5.5

According to the developers, the new model generates responses more than 30% faster than its predecessor Sonnet 5. Due to the reduction in the number of required steps and tokens, the execution of a typical task can cost up to 30% less, although the basic API rates remain the same: $2 for 1 million input tokens and $10 for 1 million output tokens, as well as $0.20 for 1 million tokens when reading from the cache.

The largest increase was recorded in coding tasks: in the Terminal-Bench 4.0 benchmark, the model scored 70.6% versus 10.3% for the previous version, and in CursorBench 4.0 - 55.5% versus 34.1%. In the comprehensive test of working scenarios GDPval-AA v2.1, the system showed a result of 1844 points, closely approaching the indicator of the flagship Opus 5.5 (1846 points). In addition, Sonnet 5.5 supports a context window of up to 1 million tokens and output of up to 128 thousand tokens.

In the updated version, the company first introduced enhanced security mechanisms in the field of cybersecurity and protection against model distillation for the Sonnet family. Access to Sonnet 5.5 is already open in the Claude chatbot, through the company's API, as well as on the Amazon Web Services, Google Cloud, and Microsoft Azure platforms.