How to Use Clean Up in iOS 27 Photos: Step-by-Step Guide

iOS 27 splits Apple's Clean Up tool into three processing modes: Auto, Fast, and High Quality, and picking the right one is the difference between an edit that holds up and one that looks obviously doctored. This guide covers how to use Clean Up in iOS 27 Photos to remove unwanted objects from iPhone photos, including which mode to trust and when to override it. Hands-on tests from Macworld and TechRadar show High Quality producing noticeably more convincing results than Fast in complex removals, from crowded street scenes to objects overlapping a person's hands (Macworld, last month).

Clean Up itself isn't new; it's been part of the Photos app since iOS 18. What changed in iOS 27 is that the High Quality option can now tap Apple's larger Foundation models for context-aware reconstruction, instead of relying only on the smaller model that has always run on-device (TechRadar, two months ago). Engadget published a walkthrough of the current tool yesterday, treating it as ready to use on compatible iPhones (Engadget, yesterday). The comparison tests cited throughout this guide, including TechRadar's, were run on the iOS 27 developer beta months earlier, so treat those specific before-and-after results as directional rather than a guarantee of what your exact device will produce.

iOS 27 Clean Up requirements: which iPhones qualify

Before opening the Photos app, confirm the basics. Engadget's guide states the upgraded tool requires an iPhone 15 Pro or later, and that anyone on an older model won't have access to it at all (Engadget, yesterday). An iOS 27-compatible device isn't enough on its own; Apple Intelligence has to be turned on too.

Apple's own support page on Clean Up requirements lists which devices generally qualify for Apple Intelligence features: iPhone 16 models plus iPhone 15 Pro and iPhone 15 Pro Max, along with iPads running an A17 Pro or M1 chip and Macs with M1 or later (Apple Support, April 2025). That page predates iOS 27, so it doesn't confirm the final compatibility list for the Fast and High Quality modes described here. Engadget's iPhone 15 Pro or later claim is the more current one to go by for iOS 27 specifically. Neither source lists the standard iPhone 15, so don't assume eligibility just because a device is a recent-ish iPhone.

Check your exact status before trying any of the steps below:

• OS version: iOS 27 installed, confirmed under Settings > General > Software Update
• Device model: iPhone 15 Pro or later, verified under Settings > General > About (Engadget, yesterday)
• Apple Intelligence: turned on under Settings > Apple Intelligence & Siri
• EU accounts: Apple Intelligence became available to EU users starting with iOS 18.4 on supported hardware, so an outdated EU install may need updating before Apple Intelligence appears at all (Apple Support, April 2025)
• Mainland China restriction: Clean Up does not currently work on iPhones purchased in mainland China. It also won't work on a supported device bought elsewhere if the phone is physically located in mainland China with an Apple Account region also set to mainland China (Apple Support, April 2025)

If the requirements above aren't met, the upgraded Clean Up controls may not appear at all. Check Apple's current support documentation for your specific model before assuming something is broken.

How to use Clean Up in iOS 27 Photos: step-by-step

Once eligibility is confirmed, the workflow for iPhone Photos AI object removal takes just a few taps.

1. Open the Photos app, select the image with the distraction you want gone, and tap Edit (Engadget, yesterday).
2. Open Tools and select Clean Up (Engadget, yesterday).
3. Look for the Auto button at the top of the screen. Leaving it on Auto lets the iPhone decide whether to run Fast or High Quality processing based on how complex the request is. Tapping it opens a menu where you can force either mode manually instead (Engadget, yesterday; TechRadar, two months ago).
4. Let Clean Up auto-highlight objects it recognizes, or circle the object manually if nothing gets flagged. Automatic detection isn't always accurate, so circling the target yourself tends to give a tighter, more reliable selection than trusting the auto-highlight (Engadget, yesterday). Keep the circle snug around the object; a loose selection that catches nearby detail can end up altering things you never meant to touch.
5. Wait for Apple Intelligence to process the edit, then review the result before moving on (Engadget, yesterday).

Fast vs. High Quality: when to override Apple Intelligence's Clean Up pick

Fast uses the smaller on-device approach described in the iOS 27 coverage and is comparable to earlier Clean Up processing from iOS 18 and iOS 26. It's built for quick, low-stakes removals rather than anything structurally complex (Engadget, yesterday; Macworld, last month). High Quality is the version that analyzes what's around the object being removed to reconstruct that space more convincingly, and that analysis takes longer to run and uses more battery than Fast does (Macworld, last month; Engadget, yesterday).

The gap between the two shows up clearly in independent testing. Macworld removed people from a cobblestone street, cranes and dockworkers from a port scene, and buildings blocking a sunset view, and found Fast left obviously manipulated results in the cited street, port, and sunset tests, while High Quality looked clean enough to pass at a glance, even reconstructing two kittens hidden behind clutter with plausible-looking paws (Macworld, last month). Tom's Guide saw a similar pattern removing a travel mug and a pair of smart glasses. Older iOS 26-era processing distorted a hand and left the glasses removal nearly unusable, while iOS 27's upgraded processing reconstructed a natural-looking hand and a believable pair of eyes instead. That glasses edit took more than 30 seconds to process, which gives a sense of the time cost for one difficult request, not a fixed benchmark for every High Quality edit (Tom's Guide, three months ago).

TechRadar's side-by-side comparison of the iOS 27 developer beta against iOS 26 adds a wrinkle worth knowing. TechRadar found that Auto sometimes appeared to switch to High Quality and produced a better result than iOS 26, but in tests where the phone stuck with Fast, the gap between iOS 27 and iOS 26 shrank close to nothing, and in a couple of cases iOS 26 actually looked cleaner (TechRadar, two months ago). That's the practical case for not always trusting Auto: it's a reasonable guess, not a guarantee it will pick the mode that produces the best result for a given photo.

Use these defaults, based on the reported examples above:

• Auto: fine for everyday shots where speed matters more than pixel-level accuracy
• High Quality (forced manually): worth selecting for people, faces, hands, or objects overlapping detailed backgrounds, where the tool has to invent realistic detail rather than smear over a flat surface
• Fast: fine for small specks, dust, or simple isolated blemishes sitting against plain backgrounds

Even High Quality has limits worth knowing before relying on it. Macworld deliberately covered part of a face with a camera and asked Clean Up to fill it back in; the High Quality result generated a face that didn't resemble the actual person underneath (Macworld, last month). Tom's Guide had a similar experience removing a baseball cap: the AI-generated result changed the subject's apparent age and hairstyle rather than revealing what was actually there (Tom's Guide, three months ago). Clean Up removes or reconstructs pixels in the selected area; it does not recover or reveal what was actually hidden behind or beneath an object. High Quality is generated fill, and that guess isn't always right.

Common Clean Up mistakes to avoid before you share an edit

After any High Quality edit, zoom in before assuming it's done. Faces, hands, text, and shadows showed up repeatedly as failure points across the cited tests, which makes them worth a second look rather than a guaranteed weak spot in every photo (Macworld, last month). TechRadar ran into this directly with one of the harder cases in its comparison: removing a person seated on a backless bench while trying to keep the rest of the scene intact. The result left a stray foot behind and erased part of the bench that was never supposed to be touched, a reminder that generated fill can look smooth at first glance while still getting details wrong (TechRadar, two months ago).

Don't assume Fast has quietly improved just because it now runs under iOS 27. TechRadar's beta testing found Fast results were sometimes no different, and occasionally slightly worse, than what the same photo produced under Clean Up on iOS 26 (TechRadar, two months ago). The clearest improvement in the cited comparisons came from the High Quality path.

TechRadar's results also came from a developer beta, not necessarily the exact build most people are installing now. Apple could have adjusted how the tool behaves before wider release, so treat any single test result, including the ones cited here, as directional rather than a guaranteed outcome on every device (TechRadar, two months ago).

What to do if a Clean Up edit looks wrong

A bad result doesn't necessarily mean the tool is broken; it usually means the selection or mode needs adjusting. Work through this sequence before giving up on a photo:

1. Tap Undo, or revert to the original in the Photos app if the edit was already saved.
2. Re-enter Clean Up and redraw the selection tighter around just the object you want gone.
3. Force High Quality from the Auto dropdown menu instead of letting the phone default to Fast.
4. Zoom in on edges, text, hands, faces, and shadows near the edit before trusting the result.
5. Stop using Clean Up on that particular image if the edit changes someone's identity or alters a meaningful detail in the scene.

That last point matters more than it sounds. A Clean Up edit is a plausible reconstruction of what could belong in that space, not a photographic record of what was actually there.

What to do next

Confirm the iPhone model and Apple Intelligence status using the checklist above before troubleshooting anything inside Clean Up itself. Once eligibility is confirmed, default to Auto for casual edits, but get in the habit of manually forcing High Quality for anything involving people, faces, or cluttered backgrounds where a convincing result actually matters.

Because Clean Up generates plausible pixels rather than recovering real ones, save any High Quality edit as a copy rather than treating it as a factual record, especially for photos where accuracy is the point rather than a clean feed.