Google recently released the Pixel 10 series smartphones and Tensor G5 processors, because the Tensor series processor is the first OEM product, and it is currently attracting attention. However, the Tensor G5 processor is indeed outstanding in some aspects, but the GPU performance is not as impressive as Google's promotion.
According to Phonearena, the Tensor G5 processor adopts the second-generation 3nm N3E process of TECHNOLOGY, which is more than the 4nm progress of the previous generation Tensor G4. Advanced processes usually represent better performance and higher efficiency. Google calls the Tensor G5 CPU speeds up by 34%, and the self-developed acceleration machine learning (ML) and artificial intelligence (AI) TPU performance is increased by 60%.
Tensor G5 CPU architecture has also changed, from one large core, three computing cores and four energy cores to one large core, five computing cores and two energy cores. Tensor G5 makes daily use of new phones more fluid, Google also provides powerful AI capabilities, equipped with a new image sensor (ISP) to enhance camera capabilities.
When posting, Google mentioned the Tensor G5 GPU in a slight way, and only stated that it had updated the IP and performed well when executing high-level games, but no details were provided. Market news says Google has changed the GPU core from the Arm Mali series to the Imagination DXT-48-1536 GPU. Google proves that the GPU lacks the hardware-accelerated ray tracing capabilities, while Apple and Qualcomm are used to develop realistic light and reflection effects.
Basic test results for network flow, the Pixel 10 Pro scored only 3,707 points in the Vulkan baseline test. This is not only far below the 26,333 points for top-level smart phones such as the Samsung Galaxy S25 Plus, but it is even lower than the 9,023 points for its predecessor Pixel 9 Pro. Such a significant regression is undoubtedly worrying, which also explains why Google has little discussion about GPU performance,
Market expectations may be related to the current low GPU operating frequency and Google has not updated it to the latest driver. Nevertheless, such low performance is still worthy of attention, especially since Google only mentioned GPU performance in a light manner, which also makes potential Pixel 10 series smart phones only hope that the GPU will improve after being updated to the latest driver.
Overall, the Google Tensor G5 processor has shown exciting progress in CPU and TPU, benefiting from the adoption of advanced TEU processes and optimized core configurations, which is expected to bring experience in daily use and AI tasks. However, the performance of GPUs is sluggish, especially the lack of hardware-accelerated optical tracking and surprising run-score test scores, casting a layer of shadow on its overall performance. Next, consumers are expected to continue to pay attention to whether Google can fully utilize the potential of the Tensor G5 processor through software updates in the future.