claude-sonnet-5-5

Description

Anthropic Claude Sonnet 5.5 — balances speed and intelligence for coding, agents, and production workloads. It accepts text and image input, has a 1,000,000-token context window, and supports up to 128,000 output tokens. Adaptive thinking is on by default and thinking tokens are billed as output; output_config.effort selects low, medium, high, xhigh, or max, with high as the default. To turn off up-front thinking, send thinking with type between_tools, which works at high effort or below; type disabled is rejected. Setting temperature, top_p, or top_k to a non-default value returns an error, as does forcing a tool call with tool_choice any or tool. Thinking blocks are tied to the model and the conversation; pass them back unchanged on later turns.