Configuring response thinking levels
Storytell lets you control the level of “thinking” applied to responses for each prompt. This setting determines how the AI processes and reasons when generating answers, so you can match the depth of reasoning to the task at hand.
Written By Patrick Intervalo
Last updated About 19 hours ago
By adjusting the thinking level, you can customize how detailed or advanced the AI’s reasoning will be based on the complexity of your prompt or the type of answer you need.

Thinking Levels
Enable or disable extended thinking
You can choose whether the AI should apply extended reasoning to your prompt.
Three thinking levels
When extended thinking is enabled, you can select from Off, Medium, or High to control how much reasoning headroom the model receives.
What each level means
Off — Standard response without extended thinking. Best for simple tasks or concise answers.
Medium — Balanced thinking for complex questions. Helpful for multi-step tasks or moderately detailed explanations.
High — Deep reasoning for challenging problems. Ideal when you need substantial context, multi-step reasoning, or thorough analysis.

Token reference
When High thinking mode is enabled, the model can use up to 32,000 tokens for reasoning and context, allowing more information to be considered when generating a response.
How it works
Click on the three-dot chat menu.

Locate the Extended Thinking button on the menu.

Click the button to choose your preferred thinking level:
Off — grayed out
Medium — highlighted in green
High — highlighted in orange
Once you click a level, the button’s color updates to show the change.
The selected level applies to the next prompt you send.
When to adjust thinking levels
Use thinking levels intentionally based on what you’re trying to accomplish:
Choose Off for quick checks, short rewrites, or straightforward questions where speed matters more than depth.
Use Medium when working through structured tasks like comparisons, summaries, or multi-step workflows.
Switch to High for complex analysis, strategic thinking, or prompts that require pulling together context from multiple sources.
You can change the thinking level at any time, making it easy to dial reasoning up or down as your task evolves.