Eight days before its launch, the first benchmark results for the A20 Pro appear in the Geekbench database. The multi-core scores are significantly higher than Apple's own announced specifications – however, the graph accurately reflects the announcement.
The iPhone 18 Pro models will be available on September 18th, and pre-orders begin tomorrow. Entries for both devices have been listed in the Geekbench database since this morning, providing a first look at the performance of the A20 Pro. For further details on what has changed compared to the previous generation, see the comparison between the iPhone 18 Pro and iPhone 17 Pro.
Key Facts at a Glance
- The iPhone 18 Pro achieves 4,719 points in single-core and 12,677 in multi-core.
- The values for the Pro Max are practically identical at 4,725 and 12,575.
- Compared to the A19 Pro, this represents an increase of approximately 21 to 23 percent in single-core performance and 27 to 28 percent in multi-core performance.
- The metal value of the seven-core graphics card is around 38 to 40 percent higher than the A19 Pro.
- The entries are unconfirmed and are based on individual measurements.
What's in the database
The listings indicate a single-core score of 4,719 and a multi-core score of 12,677 points for the iPhone 18 Pro. The Pro Max achieves 4,725 and 12,575 points – the difference is within the margin of error, as both models use the same chip.
Each test uses a six-core CPU with a clock speed of approximately 4.93 GHz. The graphics results from Metal show the full seven-core variant of the GPU, which was previously visible in a leaked chip image.
| Value | iPhone 18 Pro | iPhone 18 Pro Max |
|---|---|---|
| Single-core | 4.719 | 4.725 |
| Multi-Core | 12.677 | 12.575 |
| CPU cores | 6 | 6 |
| Clock frequency | approximately 4.93 GHz | approximately 4.93 GHz |
Where the measured values are above the announcement
Apple claims that the six-core CPU offers up to 20 percent more performance than the A19 Pro. Database entries show slightly higher figures of 21 to 23 percent for single-core tests and significantly higher figures of 27 to 28 percent for multi-core tests. Therefore, the performance increase is noticeably greater when multiple cores are under load than the manufacturer's specifications suggest.
The graphs confirm both claims: Apple cites up to 40 percent, while measurements show 38 to 40 percent compared to the six-core graphics of the A19 Pro. The new Neural Engine and the 50 percent increase in memory bandwidth are not included in these figures – Geekbench does not measure them.
What individual measurements do not show
The entries are unconfirmed and not independently verified. Anyone can submit results to the Geekbench database, and device names can be entered. Reliable conclusions require multiple runs by testers with the devices themselves.
Above all, a single test run reveals nothing about performance under sustained load. This is precisely where the interesting point lies with the Pro models of this generation, because Apple is explicitly targeting continuous performance with the vapor chamber and new chip enclosure. A benchmark that runs for just under a minute doesn't achieve this level of performance. Anyone wanting to know how the device performs after twenty minutes of gaming should wait for the tests starting September 18th.
A20 Pro: What the numbers are worth
The values serve as a first indication, and they point in a direction that is more favorable than predicted. They will only become reliable once several measurements from commercially available devices are available – and once someone measures what remains after half an hour of load.
When it comes to a smartphone chip, is peak performance more important to you, or what it still delivers after twenty minutes of continuous use? Let us know in the comments what you look for when switching.
