New Pixel 9 leaks: more disappointing benchmark results raise concerns about the Tensor G4
As the tradition dictates, the Pixel 9 line is leaking many months before its official unveiling. This time around, it's another portion of benchmark testings that reveal details about the upcoming Pixel 9. Spoiler alert: things are not looking too good for Google's 2024 handset.
The latest findings about the Pixel 9 and its Tensor G4 chipset that it's going to pack allege that it could very well be on par with the Tensor G3 – if not worse! The Tensor G3 is incorporated in the Pixel 8 family.
The reason behind these claims is that the Pixel 9 reportedly appeared in the Geekbench benchmark database. The Geekbench benchmark database is a collection of performance test results from various gadgets, including phones, that measures and compares the processing power and efficiency of these devices using standardized tests.
So, the Pixel 9 popped up in an 8GB RAM variant and in the entry, it's said that it got 1653 points for the single-core test and 3313 points multi-core.
Of course, given that the
Pixel 9 is, as said above, still months away, this whole Geekbench entry could be a fake. On the other hand, it could very well be real.
This means that Google has still plenty of time left to fine-tune the Pixel 9 and to tweak its performance.
However, I think that tweaked or not, the Pixel 9 with the Tensor G4 shouldn't record such numbers. In the worst case scenario, it should be allowed to be on par with the previous generation chipset, but not behind it.
If the Tensor G4 turns out to be a disappointment, many users will simply stick to their Tensor G3 – and even G2s – and wait for the Tensor G5.
That's because with the 2025 Pixel 10 series, Google will fully customize the Tensor G5. Instead of using Samsung Foundry to produce the 4nm chipset, the Tensor G5 will be produced by TSMC using its second-generation 3nm process node.
Mere hours ago, the Pixel 9 XL was spotted on the Geekbench database. For the single-core test, it got 1378 points and 3732 – for the multi-core. Again: take these findings with a pinch of salt, as the testing process is not clear. However, they're not impressive.
At the beginning of June, AnTuTu benchmark test results that we reported on showed that the Pixel 9 scored 1071616 points. The Pixel 9 Pro had a score of 1148452, and the Pixel 9 Pro XL had the top score among the three upcoming Pixel handsets tallying 1176410. That compares with a score of close to 900000 for the Pixel 8 and 1142984 for the Pixel 8 Pro.
The latest findings about the Pixel 9 and its Tensor G4 chipset that it's going to pack allege that it could very well be on par with the Tensor G3 – if not worse! The Tensor G3 is incorporated in the Pixel 8 family.
So, the Pixel 9 popped up in an 8GB RAM variant and in the entry, it's said that it got 1653 points for the single-core test and 3313 points multi-core.
That's eyebrow-raising to say the least, as the current Pixel 8 with the Tensor G3 is reported to record higher Geekbench numbers: 1711 single-core and 4382 multi-core.
This means that Google has still plenty of time left to fine-tune the Pixel 9 and to tweak its performance.
However, I think that tweaked or not, the Pixel 9 with the Tensor G4 shouldn't record such numbers. In the worst case scenario, it should be allowed to be on par with the previous generation chipset, but not behind it.
If the Tensor G4 turns out to be a disappointment, many users will simply stick to their Tensor G3 – and even G2s – and wait for the Tensor G5.
That's because with the 2025 Pixel 10 series, Google will fully customize the Tensor G5. Instead of using Samsung Foundry to produce the 4nm chipset, the Tensor G5 will be produced by TSMC using its second-generation 3nm process node.
The previous benchmark testings
AnTuTu scores for the Pixel 9 Pro XL, Pixel 9 Pro, and Pixel 9 | Image credit – Rozetked account at X
Mere hours ago, the Pixel 9 XL was spotted on the Geekbench database. For the single-core test, it got 1378 points and 3732 – for the multi-core. Again: take these findings with a pinch of salt, as the testing process is not clear. However, they're not impressive.
Things that are NOT allowed: