Tensor G5 exhibits middling performance alongside Pixel 10 Pro Fold on Geekbench

The Google Tensor G5's maiden Geekbench showing was somewhat lacklustre, understandable, given that it was probably an engineering sample. Now, we get to see it in action once again, this time, alongside Google's next-gen foldable, the Pixel 10 Pro Fold.

It scores 2,276 and 6,173 in Geekbench's single and multi-core tests, respectively. Geekbench's back-end reveals the chip is not throttling and can maintain a steady boost clock. While that is measurably better than its first set of scores (1,323/4,004), it is remarkably behind the current-gen market leaders, such as the Snapdragon 8 Elite (3,070/9,251) and Dimensity 9400 (2,597/8,109). It is closer to the last-gen Dimensity 9300 (2,207/7,408).

We also get a glimpse at the Tensor G5's CPU layout, consisting of one Cortex-X4 prime core (3.78 GHz), five Cortex-A725 (3.05 GHz) performance cores and two Cortex-A520 efficiency cores (2.25 GHz). And we get a glimpse at its shiny new Imagination Technologies' PowerVR DXT-48-1536 GPU. The Pixel 10 Pro Fold in question has 16 GB of RAM and runs Android 16.

Realistically speaking, the Tensor G5 could never compete with current-gen behemoths, especially with older Arm cores. Nevertheless, the transition to TSMC's 3 nm (or 5 nm, depending on who you ask) node might help with power efficiency. On the other hand, its Exynos modem might have a detrimental effect.