The benchmark numbers for the iPhone 18 Pro are in, and they tell a sharper story than the one Apple told on stage. Apple’s September slide promised a 20% faster CPU. The lab results came back at 23.5% single-core and 28.2% multi-core over the iPhone 17 Pro. Graphics jumped 40%. The chip even outran Apple’s own M5 Max on single-threaded work. None of that is the interesting part.
The interesting part is what happens twenty minutes in, when the phone is hot and the silicon starts protecting itself. That is where the 18 Pro Max pulls away from the 17 Pro Max by a margin no spec sheet advertises. And it is where the smaller 18 Pro quietly fails to keep up with its own big brother.
One housekeeping note before the tables. There is no plain “iPhone 18” this year. Apple shipped only the 18 Pro, the 18 Pro Max and the folding iPhone Duo. So the honest comparison is Pro to Pro, Max to Max, with the Duo riding the same A20 Pro chip but carrying no independent third-party benchmarks yet.
The scoreboard
Here is every number that matters, pulled from Tom’s Guide, CNMO, Notebookcheck and the UL 3DMark database. Ranges, not single points — the same A20 Pro scores differently across outlets because of iOS builds, ambient temperature and background load. Treat these as the spread you’ll actually see.
| Benchmark | 18 Pro | 18 Pro Max | 17 Pro | 17 Pro Max |
|---|---|---|---|---|
| Geekbench 6 single-core | 4,735 | 4,741 | 3,834 | — |
| Geekbench 6 multi-core | 12,800 | 12,789 | 9,988 | — |
| 3DMark Solar Bay Unlimited | 16,288 | 17,850 | 12,208 | 12,260 |
| 3DMark Steel Nomad Light | — | 3,786 | — | 2,469 |
| Geekbench Metal (GPU) | — | 64,069 | — | 45,683 |
| AnTuTu V11 | — | 3,551,073 | — | 2,275,118 |
| Wild Life Extreme stability | 62.8% | 79.4% | 61.1% | 65.2% |
Read down the single-core and multi-core rows and the 18 Pro and 18 Pro Max are the same phone. 4,735 against 4,741, 12,800 against 12,789 — that is measurement noise, not a hierarchy. The Max does not buy you a faster chip. It buys you a bigger body, and the body is the whole point.
Two footnotes on the numbers. AnTuTu’s V11 moved its GPU test to the Unreal Engine 5 renderer, so the 56% jump there is part hardware, part new engine — it is not directly comparable to V10 scores from older phones. And Geekbench Metal’s +40% (64,069 against 45,683) lines up exactly with Apple’s own graphics claim, which makes it the cleanest single piece of evidence that the GPU really did move.
The iPhone 18 Pro: same 6.3-inch footprint as the 17 Pro, a different chip underneath
What actually changed under the hood
The A20 Pro is Apple’s first 2nm phone chip, built on TSMC’s N2 process with the new gate-all-around transistor. Core count did not change — still two performance cores and four efficiency cores. What changed is the speed those cores hit: 4.93 GHz on the performance side, right at the edge of 5. The GPU grew a core, six to seven. The Neural Engine doubled, from 16 cores to a dual-16 arrangement giving 32.
The upgrade nobody put on a slide is memory. The bus widened from 64-bit to 96-bit, a 50% jump in bandwidth, paired with 12GB of LPDDR5X. That single change explains why Apple can claim +40% sustained performance while peak CPU only moved +20%. Bandwidth is what lets the chip keep feeding data to a hot GPU without stalling. It is the difference between a sprint and a distance run.
| Spec | A20 Pro (18 series) | A19 Pro (17 series) |
|---|---|---|
| Process | TSMC N2, 2nm GAAFET | 3nm (N3P) |
| CPU | 6-core, perf cores to 4.93 GHz (+20%) | 6-core, lower clocks |
| GPU | 7-core (+40% graphics, 2x FP8) | 6-core |
| Neural Engine | 32-core (dual 16), 2x AI | 16-core |
| Memory | 12GB LPDDR5X, 96-bit (+50% bandwidth) | 8/12GB LPDDR5X, 64-bit |
| Cooling | 3x vapour chamber + nanotwinned copper | vapour chamber |
| Modem | Apple C2 (uplink +50%, radio power −15%) | Apple C1X |
Peak speed was never the problem
Every flagship has been fast enough for years. Opening apps, scrolling, shooting — a 17 Pro does not stumble. The gap between the two generations only shows up under sustained load, and here the numbers split down the middle.
In the Wild Life Extreme stress test, the 18 Pro Max held 79.4% of its peak performance across twenty loops. The 17 Pro Max managed 65.2%. That is a 14.2-point gain, more than a fifth better at staying fast when it is hot. Now look at the smaller 18 Pro: 62.8% against the 17 Pro’s 61.1%. A 1.7-point gain. Practically flat.
The cooling dividend went almost entirely to the big phone. Apple tripled the vapour chamber surface area and added a nanotwinned-copper heat spreader, then laid the chip and memory side by side so the RAM stopped blocking the SoC’s escape route. In the Max’s larger chassis, that hardware has room to work. In the 6.3-inch Pro, it does not. If sustained performance is why you are upgrading, the Max is the only version where the claim holds.
| Sustained-load result | 18 Pro / Pro Max | 17 Pro / Pro Max |
|---|---|---|
| Genshin Impact, 30 min body temp | 35.3 C / ≈38 C | 45.5 C / — |
| Genshin average power draw | 2.92 W / 3.0 W | higher (Pro Max −1.9 W, −38.8%) |
| Wuthering Waves average fps | 50.65 fps | 33.2 fps (+52.6%) |
| Honkai: Star Rail frame variance | 0.2 (Pro Max) | n/a |
The Wuthering Waves result is the one to remember. It is a heavy GPU title, and the 17 Pro throttled down to 33.2 frames per second. The 18 Pro held 50.65 — a 52.6% lead that has nothing to do with a faster chip and everything to do with a chip that stays cool. Frame variance of 0.2 on the Max in Star Rail means the frame pacing is essentially a flat line. In a tropical climate like the Philippines, where the phone starts warm before you even launch a game, this is the difference you will feel.
The AI numbers, and the catch
Apple doubled the Neural Engine, so the AI benchmarks duly soared: 69,245 on the 18 Pro and 71,974 on the Max, against 49,050 and 48,549 for the 17 series. That is +41% and +48%. On a low-precision GeekBench AI run the A20 Pro’s NPU even beat Apple’s desktop M5.
Two things temper that. First, the Galaxy S26 Ultra still scores higher on the same AI test — 83,047, about 15% ahead of the 18 Pro Max. It is the one category where Android still leads. Second, and more telling: when Mashable ran an actual on-device image-generation task, the 18 Pro Max and the 17 Pro Max finished in exactly the same time. Not close. Identical.
The hardware is there. The software has not caught up to it. Today’s Apple Intelligence features are limited by model size and iOS 27’s scheduling, not by the silicon. The doubled Neural Engine is headroom for the next two years of features, not something you will feel this afternoon.
Battery, charging, and a heavier Max
Charging is the most concrete daily upgrade. The 18 series moves to a 60W wired adapter, up from 40W on the 17 Pro Max. Zero to fifty takes about fifteen minutes instead of twenty; CNET measured a real peak of 52W against 35W and called these “the fastest-charging iPhones I have ever tested.” A full charge lands in roughly 55 minutes versus 72. Wireless stayed put at 25W for both MagSafe and Qi2, so the gain is wired-only — and you need a 60W brick to see it.
Battery life moved far less than the performance did. On Tom’s Guide’s 5G web-browsing test the 18 Pro lasted 16h17m against the 17 Pro’s 15h21m, and the Max 18h27m against 17h54m. Gains of 3% to 6%. The 2nm efficiency dividend went into holding performance, not into extra hours. The Max pays for its bigger cell in weight — 249g, up 16g from the 17 Pro Max, heavy enough that Mashable listed it as a downside.
The camera: one real change, one honest caveat
The main sensor now has a variable aperture — six blades that stop down across f/1.48, f/1.8, f/2.8 and f/4.0. At f/1.48 it is the widest aperture ever fitted to an iPhone, against the fixed f/1.78 on the 17 Pro. DXOMARK scored the 18 Pro 172, second in the world, praising dynamic range and flare control.
The caveat comes from The Verge, which found the low-light gains “pretty minor” in practice. A wider aperture plus a shorter night-mode exposure (1 second against the 17 Pro Max’s 3) should mean brighter night shots; the difference on real photos was small. The variable aperture pays off more in group shots and depth control than in raw darkness.
So is it worth upgrading?
The reviewers are unusually blunt this year. The Verge called the 18 Pro “excellent and utterly iterative” and saw “no hugely compelling reason to pick an 18 Pro over a 17 Pro if you can find one on sale.” CNET, at 9.3, said owners of a 16 Pro or 17 Pro “can comfortably sit this year out.” Mashable would not rush to upgrade.
That is fair if you already hold a 17 Pro. It is the wrong framing if you are coming from anything older, or if you game hard in a hot climate. For those buyers the case is not “faster.” It is cooler, steadier, quicker to charge, and it will still be fast in 2028 when the software finally uses all that AI headroom. The Duo, meanwhile, is a genuinely new category — titanium, the thinnest iPhone when unfolded, and a crease that TechRadar found solved “with a texture.” It ships later, and no outlet has published independent Duo benchmarks yet.
The folding iPhone Duo runs the same A20 Pro, in a chassis Apple says holds +35% sustained performance over the 17 Pro
How big the jump is depends on what you already hold
Everything above compares the 18 Pro with its direct predecessor, the 17 Pro — the cleanest generational matchup and the one reviewers actually ran. But almost nobody shopping for an 18 Pro today is trading in a 17 Pro. If you are on an iPhone 12, 13, 14 or 15, the gap is not 20%. It is roughly two to three times the CPU, plus every hardware change Apple saved for later models: a 120Hz screen, USB-C, the Dynamic Island, a 48MP main camera and 60W-class charging.
| Your iPhone | Chip (year) | CPU vs 18 Pro* | Biggest things you’d gain |
|---|---|---|---|
| iPhone 12 | A14 (2020) | ~2.2x / 2.7x | 60Hz→120Hz, Lightning→USB-C, notch→Dynamic Island, 12MP→48MP, 20W→60W |
| iPhone 13 | A15 (2021) | ~2.0x / 2.3x | Lightning→USB-C, notch→Dynamic Island, 12MP→48MP |
| iPhone 14 Pro | A16 (2022) | ~1.7x / 1.8x | Lightning→USB-C, 20W→60W charging |
| iPhone 15 | A16 (2023) | ~1.8x / 1.9x | 60Hz→120Hz on the base model, faster charging |
| iPhone 15 Pro | A17 Pro (2023) | ~1.6x / 1.7x | 20W→60W charging, ~11h more video playback |
| iPhone 16 Pro | A18 Pro (2024) | ~1.4x / 1.5x | faster charging, far cooler under sustained load |
| iPhone 17 Pro | A19 Pro (2025) | ~1.27x / 1.31x | 40W→60W, sustained performance under heat |
*Single-core / multi-core multiples, based on Geekbench 7 cross-generation data (browser.geekbench.com). The 18 Pro is not in the official GB7 chart yet, so its reference score comes from Tom’s Guide’s lab.
Those felt upgrades matter as much as the multiples. The base iPhone 12, 13, 14, 15 and 16 all run a 60Hz screen — the non-Pro model only got 120Hz with the iPhone 17. Lightning ends at the 14; USB-C arrives with the 15. The 48MP main camera starts at the 14 Pro and reaches the base model at the 15. Each is a daily, visible change with nothing to do with a benchmark, and for anyone on a three- or four-year-old phone they outweigh the chip.
When the iPhone 18 family reaches the Philippines, you can compare every capacity and colour, check current prices and read local buyer reviews in one place on Lazada. Browse the full range and find the version that fits how you actually use your phone.
Sources
- Tom’s Guide — iPhone 18 Pro and Pro Max benchmarks and review (4.5/5)
- CNMO / Sina Tech — A20 Pro score roundup, cross-checked against Tom’s Guide data
- Notebookcheck — A20 Pro Geekbench Metal and architecture breakdown
- Geekbench Browser (browser.geekbench.com) — cross-generation Geekbench 7 scores, iPhone 12 to 17 Pro Max; Apple Tech Specs pages for display, connector, camera and charging by generation
- The Verge — iPhone 18 Pro review (8/10); CNET — iPhone 18 Pro / Pro Max review (9.3/10); Mashable — iPhone 18 Pro Max review (4.25/5)
- Apple Newsroom — iPhone 18 Pro / Pro Max and iPhone Duo announcements, 9 September 2026
- 凰家评测 (ifeng) — gaming thermal and frame-rate testing, 18 September 2026