When should I use Instant, Thinking, and Custom modes?

Phi offers different chat modes depending on what you need:

• Instant: Fast replies for general questions and conversation.
• Thinking: For complex questions that require logic, analysis, or deeper reasoning.
• Custom: For when you want to select a specific AI model and control the thinking effort.

What is Thinking Effort?

In Thinking and Custom modes, you can adjust the "Thinking Effort" if the selected AI model supports reasoning. This controls how much effort the AI puts into forming a response:

• Quick: Faster responses, less Energy used
• Standard: Balanced speed and depth
• Deep: More thorough reasoning, takes longer and uses more Energy

Higher effort means better reasoning on complex tasks, but also increases response time and Energy consumption.