The Tensor G6 Paradox: Efficiency Gains Papering Over a Multi-Core Regression
Google pitched Tensor G6 as its biggest under-the-hood upgrade in years, claiming roughly 20% better power efficiency, 25% faster web browsing, and 15% faster app loading versus Tensor G5, plus a Gemini Nano 4 model that runs up to 3.5x faster on-device while using 3.5x less energy[1]. Leaked Geekbench 7 results on the Pixel 11 Pro XL complicate that pitch: single-core climbed a modest 7% to 2,112, while multi-core actually fell about 12% to 5,196 versus Tensor G5, a consequence of Google's move to a seven-core CPU design[2][3]. GPU compute jumped roughly 80% to 7,757 points, a genuine leap, but it still lands at only about a third the speed of Qualcomm's Snapdragon 8 Elite Gen 5[2]. Community reaction to the leak seized on the multi-core dip as proof the efficiency story doesn't fully match the marketing, with extensive debate over whether pre-release firmware and drivers are masking the final chip's real performance - a question that will only be settled once retail units are benchmarked.


