KNOWLEDGE BASE
Log In    |    Knowledge Base    |    4D Home
Tech Tip: Enhancing Complex AI Outputs with Reasoning Effort
PRODUCT: 4D | VERSION: 21 | PLATFORM: Mac & Win
Published On: February 23, 2026
For complex tasks when using AI, consider using the reasoning_effort attribute available in 4D AIKit.

Some AI models support reasoning levels that control how much effort the model spends analyzing a problem before generating a response. Adjusting this parameter can significantly improve result quality for advanced or multi-step tasks.

The reasoning_effort attribute supports three levels:

  • "low": Faster responses with lighter reasoning

  • "medium": Balanced performance and reasoning depth

  • "high": Deeper analysis for complex or technical problems

Example:

$params:={model: "gpt-5"; reasoning_effort: "high"}

$response:=chat.completions.create($messages; $params)