An Anthropic model for responsive coding, long-document analysis, and text-and-image understanding with adaptive thinking.
Use anthropic/claude-sonnet-5-5 through the sjolt LLM API.
anthropic/claude-sonnet-5-5Token rates
anthropic/claude-sonnet-5-5Similar Models
An Anthropic model for careful coding, long-document analysis, and visual understanding with optional adaptive thinking.
An Anthropic model for coding, long-context knowledge work, and visual understanding with adaptive thinking.
A fast DeepSeek model for text generation and reasoning tasks.
Model overview
Build coding assistants, connect long documents, and interpret images with adaptive thinking through native Anthropic Messages.
Provide the implementation, expected behavior, and acceptance criteria. Ask for a focused diagnosis or change, then return client-side tool results with the original conversation blocks intact.
Use up to 1,000,000 context tokens for source files, specifications, documents, and conversation history. Label sources and ask for evidence supporting each conclusion.
Attach screenshots or diagrams and describe what you want to understand. Responses are text. Page illustrations are conceptual artwork, not images generated by this language model.
Model characteristics
Up to 1,000,000 context tokens and 128,000 output tokens. max_tokens includes thinking and the final response.
Adaptive thinking is enabled by default. Select low, medium, high, xhigh, max through output_config.effort; the default is high.
Set thinking.type to between_tools to skip up-front thinking while retaining thinking between tool calls. This mode accepts low, medium, high effort.
Send model, messages, and max_tokens to /v1/messages, with system instructions in the top-level system field. Read text blocks from content and preserve signed thinking blocks in API tool conversations.
State the goal, constraints, relevant evidence, and expected output format. Keep source material separate from instructions and maintain append-only history when reusing native thinking blocks.
FAQ
Send model anthropic/claude-sonnet-5-5 to /v1/messages with messages and max_tokens. Public API requests can use native streaming.
Omit thinking or set thinking.type to adaptive, and select low, medium, high, xhigh, max effort. The default is high. To skip up-front thinking, use between_tools with low, medium, high effort.
Set max_tokens to a positive integer up to 128,000. Thinking and final output share this allowance within the 1,000,000-token context window.
Send text and image attachments, choose the thinking effort, and continue the visible conversation with follow-up questions. The Playground uses adaptive thinking and the model returns text.
Preserve the original signed thinking and tool blocks with their conversation history, then append tool results and the next turn. Use automatic tool selection and clear tool schemas.