iPhone 18 Pro Max in performance tests: A20 Pro's AI power stands out
While the A20 Pro in the iPhone 18 Pro Max achieved high scores in synthetic tests, the most significant difference was seen in GPU-accelerated AI workloads.
12punto
The A20 Pro processor used in Apple's iPhone 18 Pro Max model has been measured in various synthetic performance tests. CPU and GPU performance were evaluated with Geekbench 6, AI workloads with Geekbench AI, and ray tracing performance with 3DMark Solar Bay Extreme Unlimited.
The test results show that the A20 Pro reaches a high level in classic processor performance; the most striking difference occurs when the GPU is activated for AI operations.
TEST RESULTS
In the Geekbench CPU test, the iPhone 18 Pro Max scored 4,045 in single-core and 11,554 in multi-core. These results point to a strong profile, especially in interface responsiveness, app launches, and single-core intensive tasks. The multi-core result is significant for tasks that can run in parallel, such as video processing, code compilation, and intensive calculations.
| Test | iPhone 18 Pro Max result |
|---|---|
| Geekbench CPU - Single-core | 4,045 |
| Geekbench CPU - Multi-core | 11,554 |
| Geekbench GPU - Metal | 62,689 |
| Geekbench AI CPU - Single Precision | 6,792 |
| Geekbench AI CPU - Half Precision | 12,199 |
| Geekbench AI CPU - Quantized | 9,030 |
| Geekbench AI GPU - Single Precision | 12,825 |
| Geekbench AI GPU - Half Precision | 24,323 |
| Geekbench AI GPU - Quantized | 27,142 |
| 3DMark Solar Bay Extreme Unlimited | 3,276 |
| 3DMark FPS range | 21-83 FPS |
| 3DMark average FPS | 22.9 FPS |
| Battery during 3DMark | 85% › 85% |
On the GPU side, a score of 62,689 was achieved in the Geekbench Metal test. This score reflects not only gaming performance but also general computational capacity in GPU-accelerated tasks such as image processing, computational photography, computer vision, and machine learning.
THE GPU EFFECT IN AI
Geekbench AI results revealed that the A20 Pro is significantly accelerated by the GPU in AI workloads. When using the CPU, 6,792 points were measured in Single Precision, 12,199 in Half Precision, and 9,030 in the Quantized test.
When the same AI workloads were run on the GPU, the scores rose to 12,825 in Single Precision, 24,323 in Half Precision, and 27,142 in the Quantized test. Accordingly, while an increase of approximately 89 percent was seen in Single Precision, the Half Precision result nearly doubled. The largest difference was recorded in the Quantized test, which reached approximately three times the CPU result.
These results show that running the correct AI workload on the appropriate hardware unit can make a serious difference in performance. However, it should be remembered that synthetic tests do not determine the daily usage experience on their own. Real-world performance may vary depending on the application used, the AI model, data type, framework, and Apple's software optimizations.
In the 3DMark Solar Bay Extreme Unlimited test, the device scored 3,276. While the FPS range was measured as 21-83 during the test, the average FPS was 22.9. The fact that the battery level remained the same at 85% throughout the test was also among the shared results.
The overall picture shows that the A20 Pro in the iPhone 18 Pro Max delivers strong results on both the CPU and GPU sides; the most remarkable area is the GPU-supported AI performance. Especially the jump in the Quantized AI test highlights the importance of hardware acceleration for on-device AI operations.