Claude Sonnet 5.5

Supported Functionality

Item	Specification
Input	Text, images
Output	Text
Context	1,000,000 tokens
Max Output	128,000 tokens; up to 300,000 tokens in Message Batches API beta
Vision	✓ Supported
Function Calling	✓ Supported with automatic tool selection; forcing a specific tool is not supported

Description

Claude Sonnet 5.5 is a general-purpose language model released by Anthropic on September 28, 2026. Its official Claude API identifier is claude-sonnet-5-5. It accepts text and images, produces text, provides a one-million-token context window with up to 128,000 standard output tokens, and has a reliable knowledge cutoff of June 2026.

Anthropic positions the model as the best combination of speed and intelligence for everyday development, agent orchestration, and professional work. It supports Adaptive Thinking with high as the default effort, while comparative latency is Fast. Vision and tool use help it maintain long-running task state, adapt tools, and coordinate multiple execution steps.

Key Capabilities

• Everyday Software Development: Handles implementation, defect repair, refactoring, tests, code explanation, and technical documentation.
• Agent Orchestration: Plans multistep workflows, delegates subtasks, calls tools, and reconciles or verifies execution results.
• Professional Knowledge Work: Synthesizes multiple sources and complex requirements into reports, proposals, analysis, and decisions.
• Million-Token Analysis: Connects large code, long documents, project history, and tool results within one task.
• Vision: Analyzes screenshots, charts, interfaces, and document images as evidence for subsequent action.
• Adaptive Thinking: Adjusts reasoning to task difficulty and uses effort to control depth.
• Tools and Computer Operation: Uses client or server tools for retrieval, code execution, files, and interface workflows.

Technical Strengths

Feature	Benefit
Speed–Intelligence Balance	Combines strong reasoning with Fast comparative latency for everyday development and agent work.
1M Context Window	Holds large repositories, document collections, project history, and long-running task state.
128K Standard Output	Supports complete code changes, extensive analysis, and large structured documents.
Adaptive Thinking	Runs at high effort by default and adjusts reasoning depth for different tasks.
Mid-Conversation Control	Adapts system direction or tools during ongoing work as goals and environments change.
Vision and Tool Coordination	Interprets a screenshot, interface, or document before using tools to act and verify.

Frequently Asked Questions

How should I set effort for Claude Sonnet 5.5?

The default high level is a strong starting point for complex development and agent tasks. Test lower levels for well-scoped everyday work and raise effort when deeper reasoning is required. Compare accuracy, tool success, and observed latency on the same tasks instead of fixing every request at the highest setting.

Why can Claude Sonnet 5.5 not be forced to call a specific tool?

The model rejects tool_choice values of any or a named tool. Use the default auto behavior with strict tool use, or Structured Outputs when the final response requires a fixed schema, and state the tool's application conditions clearly in the prompt.

How should I use the million-token context window?

Organize inputs by source, module, chronology, or task stage and distinguish instructions, facts, and background. Large repositories and document collections still benefit from search, file selection, staged summaries, and validation so irrelevant content does not distract from key evidence.

Should I choose Claude Sonnet 5.5, Opus 5.5, or Haiku 5.5?

Start with Sonnet 5.5 for everyday development, agent orchestration, and most professional work, and use Opus 5.5 for longer-running, more complex engineering and knowledge tasks. Haiku 5.5 fits high-volume classification, extraction, routing, and well-scoped subagent work. Select with the same evaluation set for quality and stability.

How do I call Claude Sonnet 5.5 on LinkModel?

Claude Sonnet 5.5 is available on LinkModel. Open its live model page, copy the published model ID, and submit text or image input with the current request structure. Use the live page for the platform route, context limits, tool support, and thinking options.

How do I validate Claude Sonnet 5.5 on LinkModel?

Complete a basic text request with the model ID shown on LinkModel, then test image input, long context, structured results, multistep tool calls, and agent orchestration. Record response fields, thinking blocks, tool recovery, truncation, and processing time, and confirm which options LinkModel currently exposes for effort, between_tools, Computer Use, and Batch.