While the Pixel 6 series may be challenging the latest iPhones, benchmarks suggest that Google’s Tensor chip still has a long way to go before it can catch up with Apple’s A15 Bionic chip or even last year’s A14 Bionic SoC. Google made the Tensor chip with some help from Samsung but made some tweaks to shed the Exynos shadow, especially when it comes to AI and Machine Learning chops.

The Tensor chip packs two high-performance Cortex X1 cores, a quartet of Cortex-A76 cores, and an equal number of Cortex-A55 cores for less demanding tasks. And to rise above the rivals under Exynos, Snapdragon, and Bionic A-series branding, Google played around with the TPU to accelerate on-device AI tasks such as real-time translation and intelligently mixing commands with dictation while voice typing. But it appears that there is a healthy performance gap between Google and Apple in areas where it matters — raw processing and graphics prowess.

Related: Pixel 6 & Pixel 6 Pro Specs: Google Tensor, 50MP Camera, And More

AnandTech put the Tensor chip through its paces with a series of benchmark runs against the Qualcomm Snapdragon 888, Samsung’s top-of-the-line Exynos 2100, and Apple’s A15 chip, among others. Running the Standard Performance Evaluation Corporation (SPEC) CPU tests revealed that the Pixel 6’s tensor chip is far behind the A15 Bionic inside the iPhone 13 series phones and proves to be a mixed bag against the Qualcomm and Samsung’s top-tier chips as well. A similar tale comes to the forefront when Tensor’s power efficiency is assessed.

Apple Maintains A Safe Lead Over All Its Rivals

Google Tensor Is Far Behind Apple's chips

It appears that the Tensor’s X1 cores have a throttling issue, although it is not as severe as noticed on the Exynos 2100. According to some popular synthetic benchmarking platforms, Google’s offering scored lower than Snapdragon 888 and Exynos 2100 on single-core and multi-core tests. The performance gap with Apple’s A14 and A15 chips is a lot wider. The Geekbench multi-core tally of Tensor was roughly 30 percent lower than the A14 Bionic chip fitted inside the iPhone 12, while the A15 Bionic raced ahead with a lead of over 66 percent. The single-core tally of Tensor’s Geekbench run also paints a similar story against Apple’s smartphone chips.

In the graphics-intensive 3DMark Wild Life Unlimited, Basemark GPU 1.2 (Medium at 1440, Off-Screen), GFXBench Aztec Ruins and Manhattan 3.1 tests, Google's Tensor could barely catch up with its Qualcomm and Samsung rivals. And when pitted against Apple’s A14 and A15, it was far behind. Google's chip's sustained frame rate per second output on the GFXBench Manhattan 3.1 test was half that of Apple’s A15 Bionic. At the same time, the gap with A14 Bionic was in the ballpark of 40 percent. While the Pixel 6 duo is not a slow phone by any stretch of the imagination, the Tensor still has a long way to go before it can challenge Apple’s A-series chips. Google has reportedly started working on the second-generation Tensor chip already, and it remains to be seen whether it can take a giant performance leap with its second attempt.

Next: Why Google's Pixel 6 Is The Best Phone You Can Buy For Under $600

Source: AnandTech