What is it for in our workflow
This mode is designed to resolve those classic limitations that force us to make infinite adjustments:
- Fidelity to the prompt and composition: reduces failures when we ask for complex spatial relationships or specific shots (such as full-body views or wide architectural shots), by readjusting the volumes.
- Removing unwanted elements: AI often inserts details out of context (for example, applying organic textures to smooth surfaces). This mode recognizes the inconsistency and cleans up the layout.
- Style consistency: if we are trying to replicate the mood of a specific photographer or designer and the basic model “misses the mark” on colors or lights, Thinking Mode analyzes the gap and forces the image to align with the required style.
How it works and how to use it
Thinking Mode is not a command to be inserted in the initial prompt, but a tool for correction and refinement afterwards. Here’s how to activate it:
- You must use the official web version (Midjourney Alpha).
- It works exclusively on images generated with V8.1 and V8.2 models (it is not retroactive to older V7 or V6 generations).
- Select the image you just created by opening the details window (lightbox).
- In the bottom right corner, click on the “Rerun (Thinking)” button.
The system will take a few more moments to process the information, returning four new, significantly cleaner variants.
The trick of the trade
This tool is iterative! If the first correction is close but not yet perfect, you can click “Rerun (Thinking)” again on the new image. The AI will dig deeper and deeper into your concept to align exactly with what you had in mind.
In conclusion
Tools like this remind us why we embrace a hybrid workflow. AI does not work automatically on its own, we are the ones who have to guide and correct it. Thinking Mode gives us back precious time, allowing us to save excellent concepts that we would previously have discarded due to a minor flaw, and giving us total control over the artistic direction. Until next time viewing!
The weekly column “Architectural Prompting” is edited by experts Luciana Mastrolia, Giovanna Panucci and Andrea Tinazzo
>> If you are interested in these topics, also sign up to the free Linkedin Newsletter AI & Design for Technicians, we’ll talk about it here!
