If you make covers, mockups or social graphics for digital products, you know how often an image generator nails the mood but fumbles the details: a missing object, an odd hand, or garbled text on a title. On October 8, 2026, Midjourney began testing a feature aimed at exactly that problem. It’s called Thinking Mode.
What Thinking Mode does
According to Midjourney’s changelog, Thinking Mode lets you rerun a job made with version 8.2 (standard images and edits). Midjourney compares the image against your prompt, looks for where it fell short (objects, layout, anatomy or text), and then generates again. The team says that in its own testing the mode improved prompt accuracy, typography and coherence. The announcement doesn’t include any benchmark figures.
To try it, open a job’s lightbox on the Alpha website (alpha.midjourney.com) and click “Rerun (Thinking)” in the lower-right corner. Midjourney calls this an early version and is asking for feedback on its Discord.
Why it matters for digital product creators
Text and layout errors are often what make an AI image unusable for an ebook cover, a course thumbnail or a product mockup. A pass that checks the result against your prompt could mean fewer rerolls. Since this is still a test, though, plan to judge results yourself and keep doing your usual edits in a design tool.
The same update added shared folders, also an early version. You can invite collaborators who generate into a folder or viewers who can only browse it, which may help if you work with a designer or VA. Midjourney notes that sharing doesn’t change image privacy: non-stealth images can still show up publicly.
Costs and limitations
- Midjourney’s posts don’t say which plans include Thinking Mode or whether a rerun uses extra GPU time.
- Current plans run from Basic at $10/month to Mega at $120/month, with annual discounts. Only Pro ($60/month) and Mega include Stealth Mode.
- Per Midjourney’s terms, companies (or their employees) with more than $1 million in annual gross revenue need a Pro or Mega plan to own their assets.
- It’s an Alpha-site test, so the feature may change or be released in a different form.
Copyright and disclosure
Midjourney’s terms say you own the assets you create to the extent the law allows. That doesn’t guarantee copyright protection, however. The U.S. Copyright Office’s January 2025 report concluded that purely AI-generated material isn’t copyrightable and that prompts alone aren’t enough control to make you the author. Your own creative selection, arrangement and modifications can be protected. If you sell templates or graphics, consider adding meaningful human edits and following any AI disclosure rules on the marketplaces you sell on. This is general information, not legal advice.
Sources
- Midjourney: Testing Thinking Mode
- Midjourney: Alpha Changelog – 10/7/26
- Midjourney Docs: Comparing Midjourney Plans
- Midjourney Docs: Terms of Service
- U.S. Copyright Office: Copyright and Artificial Intelligence, Part 2: Copyrightability
Featured image: Photo: Rawpixel (CC0)