Thinking

3 models

The default mode for most development work. Uses 3 models with medium thinking effort for proper planning and execution. Ideal for feature implementation, debugging, refactoring, and analysis. Claude Sonnet 4.5 consensus.

Try Now

Example Use Cases

Changes across multiple files

Implementing new features or components

Debugging issues or fixing bugs

Refactoring code structure

Writing or updating tests

Questions requiring code or documentation reading

Task involves any uncertainty or decision-making

Needs understanding of code context

Requires any form of analysis or reasoning

Technical Specifications

Model Count

3 models

Thinking Effort

Medium

Thinking Budget

2,048 tokens

Ensembler

Claude Sonnet 4.5

How It Works

Multi-Model Consensus

This mode uses 3 AI models working in parallel to analyze your request. Their responses are then synthesized by Claude Sonnet 4.5 to produce the final answer. This approach improves accuracy by leveraging multiple perspectives and cross-validating conclusions.