The first reasoning model from Anthropic. Great improvements, mainly on coding and logic, over Claude 3.5 Sonnet and 3.7 Sonnet.