Is one day enough on 100% with iPhone 18 Pro?

A mixed day is enough. Two 256GB lab units from 8:00 to 23:00: the US eSIM still had about 18%, the mainland unit about 14%. Screen-on was about 7 hours. Heavy short-form plus gaming plus 5G navigation, and you will want a socket around 9 p.m.

Apple says 36 hours of video playback—why does two hours of scrolling drop a chunk?

36 hours is the lab claim for a pure-eSIM model, local video, and a fixed lower brightness—not a short-form feed. In the same 4 hours, local 1080p dropped about 12%; short-form dropped about 15% in only 2 hours (worse when scaled). Mainland China and Hong Kong have a SIM tray; do not paste 36 / 45 hours onto those phones.

Do photos or video drain more?

Single stills are not scary. In a 90-minute walk with 4K clips, variable-aperture preview, and bursts, the US unit dropped about 16%. Leaving the camera open with Pro Controls and the histogram costs more than tapping the shutter.

How much does an hour of gaming drop? Will it thermal-throttle?

About 21%–22% at 60 fps under high load. The back gets warm, but we saw no obvious frame drop in 60 minutes. That lines up with the 17 Pro vapor chamber at about 3× and sustained performance up to about +40%. Do not multiply 21%×3 for a 3-hour session—heat bends the curve.

How much more does 5G cost versus Wi-Fi?

Same apps, same daily rhythm: all-day 5G dropped about 8 to 12 percentage points more than all-day Wi-Fi. When US mmWave is unstable, hunting a tower costs extra. Weak signal hurts more than full-bar 5G.

Can launch-day numbers be the final score?

No. Out-of-box iCloud, Photos, and Mail indexing, plus iOS 27 first-day background work, make standby and the first few hours look ugly. We wrote on September 15: wait a full 48 hours. The scene slices here ran after sync settled; the mixed day still carries a bit of day-one dirt.

iPhone 18 Pro battery life test: how long does 100% last? Video, camera, gaming, and one 5G day of data

TL;DR · six lines first

A mixed day lasts: 8:00 to 23:00, US eSIM still about 18%, mainland about 14%; screen-on about 7 hours
36 hours is local video, not short-form: 4 hours of local 1080p dropped about 12%; a short-form feed dropped about 15% in 2 hours
Camera drain is preview and 4K: a 90-minute walk with clips dropped about 16%–17%; a single shutter press barely shows
Gaming ~21% in 60 minutes: the back gets hot, but no obvious throttle in an hour; do not estimate an all-nighter as 21%×3
All-day 5G costs about 8–12 percentage points more than Wi-Fi; hunting a weak tower hurts more than full bars
Day-one numbers look uglier: indexing and iCloud are still running. Judge endurance after at least 48 hours

On September 18, pre-ordered iPhone 18 Pro units started landing. We already unpacked the stickers in worth it after the hike: US $1199, mainland 9999 yuan, video playback up to about 36 hours on the pure-eSIM claim, Pro Max about 45 hours. How to hand in the old phone is in the trade-in process. Comments then switched to the plainer question: from a full charge, can you get home? How much do video, photos, games, and 5G each take?

This is a Dev Journal two-phone, one-day slice, not a hundred-unit climate-chamber report. The 36-hour claim is there to match the exam Apple set; a real day answers “what percent is left at 11 p.m.” Mainland China and Hong Kong keep a physical SIM, so the battery bay is a notch smaller than US and Japan pure eSIM—do not paste the US 36 / 45 hours onto a mainland unit.

18%

US mixed day left at 23:00

Mixed-day screen-on (about)

36h

Official local video (eSIM)

How we tested: two phones, four scenes, what we skipped

Both units are 256GB iPhone 18 Pro, iOS 27, battery health at 100%. One US Glacier, pure eSIM; one mainland same color, physical SIM + eSIM. Brightness locked at 50% for the scene slices; the mixed day used Auto-Brightness outdoors. Always On and 120Hz stayed on. Location stayed on system defaults.

Unit	What it is for	What it does not represent
US 18 Pro · eSIM	The official 36-hour battery bay; US 5G / mmWave	Not Pro Max, and not Hong Kong or Japan carrier radio
Mainland 18 Pro · physical SIM	How long a tray model lasts in a day; mainland 5G	Not a “Hong Kong bargain” or a US gray-market import

Unbox day syncs Photos, Mail, and the iCloud keychain. We ran the scene slices after the sync bar stopped jumping; the mixed day was a real outing, and we admit the first hours still had indexing. Why iOS 27 public-release day one looks ugly is in the drain and heat field test—a new phone has the same problem.

What we deliberately skipped

We did not dim to Apple-lab “looks a bit dark” brightness just to chase 36 hours
No teardown ammeter; numbers are from system Battery and a stopwatch
No same-day 17 Pro duel—some old phones already went to trade-in, and health would pollute the sample
No Pro Max. 45 hours and “about 30 hours under the new usage model” are another machine

First, the claim: what 36 hours actually is

Apple wrote video playback, not “Recs, replies, and one camera session” added together. On the launch spec sheet, a pure-eSIM 18 Pro video claim is about 36 hours, wired about 15 minutes to 50%; Pro Max video about 45 hours, and about 30 hours full charge under the new usage model. The C2 modem is advertised at about 15% lower power; the vapor chamber is about 3× the 17 Pro area. The full spec table is still in the September 10 piece.

Lab video is usually: a local file, mid-to-low brightness, radios off or pinned, a clean background. Your day is: 5G hunt, push, Always On, short-form decode, camera ISP, the occasional game. A 2× gap is not “fake marketing.” It is a different exam.

US/Japan vs mainland China and Hong Kong

The US and Japan give the tray space to the battery; advertised video is about 2 hours longer. Mainland China and Hong Kong keep a physical SIM—do not paste 36 / 45. On the same mixed day, the mainland unit dropped a few more points than the US one. The tables later match that; the gap is not “half a battery class.”

Video: local playback vs short-form

“Watching video” is at least two questions. A local file is close to Apple’s claim; a feed is the thing that burns your eyes at night.

Local 1080p

The same 1080p file looped 4 hours, 50% brightness, cellular off, Wi-Fi off, push left on.

	US eSIM	Mainland
4-hour drain	about 12%	about 13.5%
Linear extrapolate	about 33 hours	about 30 hours

Extrapolation is not a measured 33 hours. Change brightness, bitrate, or background work and the number moves. It only says: on Apple’s exam, 18 Pro can walk up to that 36-hour class, and mainland is probably two or three hours shorter—same direction as “the tray yields battery space.”

Short-form feed

Same brightness, cellular on, normal push, 2 hours of portrait short-form (decode, recommendation requests, comment scrolling). US dropped about 15%, mainland about 16%. Scale that to 4 hours and you are a bit over 30%—the local file only dropped 12% in the same window. The gap is always-on radio, bitrate spikes, touch, and recommendation refresh, not a suddenly broken panel.

So Keynote’s 36 hours and “why did two hours of scrolling drop so much” can both be true. Sit a feed down for a lab exam and the grade will look bad.

An iPhone on a white desk scrolling an image feed, standing in for the iPhone 18 Pro short-form and social battery test — Short-form and recommendation feeds are an endurance exam—not Apple’s claimed “video playback.”

Camera: the shutter is cheap; preview and 4K are not

18 Pro’s main camera is 48MP Fusion, the first variable aperture; the viewfinder can also open Pro Controls and a histogram. Those look great on stage. In Battery they are ISP plus a screen that stays on.

90-minute walk: about 180 stills, two ~2-minute 4K clips in the middle, more live preview than lock screen. US dropped about 16%, mainland about 17%. Control: the same 90 minutes of shutter taps, little 4K, frequent lock—drain lands near 8%.

A single still: barely visible; do not use it to judge “camera drain”
Variable-aperture preview + histogram left on: screen and ISP both spin; more expensive than Auto
4K clips: the real camera-day hog; travel vlogs count as video, not “how many photos”

Mainland launch units still lack Apple Reference Image; that does not save much power—the certification signature is not the endurance driver. For the camera spec itself, go back to the spec table.

Gaming: ~21% in an hour, warm, but it did not fold

High-load 60 fps, quality on the high side, cellular on, 60 minutes straight. US dropped about 21%, mainland about 22%. The back got clearly warm—this is not the same heat as iOS 27 indexing: GPU + vapor chamber doing the job.

No stair-step frame drop the eye could catch in 60 minutes. That lines up with the keynote “sustained performance up to about +40%,” but you cannot extrapolate to 63% over 3 hours. Once heat builds, the system trades smoothness against power; there is no fan, so the back is the heatsink.

Foldable Ultra endurance and hinge are a different exam; see the Ultra launch breakdown. The slab 18 Pro’s edge is one-hand use and a mature sustained curve—not two cover displays lit at once.

Mixed 5G day: 8:00 to 23:00

This is the number most people actually want. Both phones left at 8:00 and came home at 23:00, with messages, maps, about 80 minutes of short-form, a short photo stretch, one ~25-minute game, browsing, and mail. Always On stayed on. One unit ran 5G all day (the US one had a stretch of weak mmWave); the mainland unit ran mainland 5G; a control day put the US unit on almost all-day Wi-Fi.

Day type	Left at 23:00	Screen-on (about)	How to read it
US · 5G mixed	about 18%	about 7h10m	Ordinary work + a scroll + a shoot; you get home
Mainland · 5G mixed	about 14%	about 7h05m	Same rhythm, a bit hungrier; not “half a generation”
US · near all-day Wi-Fi	about 28%	about 7h	5G that day cost about 10 extra points

Heavy users who push short-form to 3 hours, add another hour of gaming, and keep navigation up will see the 20% warning around 9 p.m. That is not “18 Pro endurance is bad.” The exam changed from mixed day to high-pressure day. Start at 100%, and an ordinary day lasts until bedtime; a high-pressure day wants a top-up after dinner.

Overnight: first night out of the box, both units dropped about 9%–10% (indexing unfinished); a later control night after sync settled was about 4%–5%. Same class of dirty data as iOS 27’s “first night extra 8%–15%.” Do not use it to judge battery health.

What actually steals power is not “a smaller battery”

In Settings → Battery, the mixed-day leaders are boring, and stable:

Display: Always On + 120Hz + short-form still finishes first
System Services / Cellular: 5G and weak-signal tower hunt; US mmWave edge is especially obvious
Camera: only jumps when you actually shoot or record; preview costs more than the shutter
Siri & Dictation / Apple Intelligence: regions where the US unit can use them take a slice; mainland and the EU do not get next-generation Siri AI at launch, so this line should not take the “resident large model” blame

To save power, first drop high refresh where you do not need it, 5G on a weak cell, and Camera left open in the background—not a suspicion that Apple shipped a smaller cell. C2’s advertised 15% cut assumes a clean network; in a basement or an elevator, even a good baseband is still hunting.

Top-up: can 15 minutes really get you to 50%?

On the US unit with official-rated wired power, from about 8% plugged in, 15 minutes to 49%–52%, which matches “about 15 minutes to 50%.” Mainland charging policy is a bit more conservative by region; the same window landed around 46%–50%. That is emergency top-up, not a habit of fifteen-minute shallow charges every day.

Pro Max also advertises “about 7 hours of video from 5 minutes.” That is another phone and another lab exam. Do not take a Max slice to a group chat for 18 Pro.

FAQ

From 100%, is one day enough?

A mixed day, yes. Our 8:00–23:00: US about 18% left, mainland about 14%, screen-on about 7 hours. Stack short-form, gaming, and navigation, and pack a battery after dinner.

Why is it not 36 hours?

36 hours is the local-video claim on eSIM models. Short-form, 5G, camera, and Always On are not on that exam. Mainland China and Hong Kong should not use this number.

Will photos alone drain the battery fast?

Shutter taps alone will not. Leaving preview and 4K open will. 90 minutes of “actually shooting” is about 16%.

Is gaming an endurance killer?

About a fifth of the battery in an hour—heavy, not a collapse. Heat is expected; do not linear-extrapolate a multi-hour session.

Should I turn off 5G to save battery?

If the signal is good and you need the speed, leave it on. Weak cells, basements, high-speed-rail edges: LTE or Wi-Fi is usually cheaper than forcing 5G. All-day gap is about 8–12 percentage points.

Can launch day be the verdict?

Not as a final score. Read Battery for at least a full 48 hours. Developers should not treat unbox day as an app-drain baseline—put compiles and Instruments on a Cloud Mac, and leave the phone for a clean day.

Should I just get the Pro Max?

If you need “two travel days without a charger” or that official 45-hour claim, look at Max. If you want one hand and a day that already works, the 18 Pro mixed-day leftover says it is not the short-endurance SKU.

Close

Three lines: iPhone 18 Pro can go from morning to bedtime on an ordinary 5G mixed day—US about 18% left at 11 p.m., mainland about 14%. Apple’s 36 hours is a local-video exam; short-form, camera preview, and gaming are a different paper. Tray models lose a little, not “half a generation of battery.” Day-one indexing makes standby look ugly; wait 48 hours before you hate this cell.

How long 100% lasts depends on which exam you sat. Count how much of your day is Recs, camera, and 5G before you swap to Max for two or three points—or hand in the old phone for credit.