MediaTek Dimensity 9600 Pro Leak Suggests 5GHz Clock Speed, High Benchmark Scores

MediaTek’s next-generation flagship chipset, the Dimensity 9600 Pro, is rumoured to rival the purported Snapdragon 8 Elite Gen 6 Pro. While the Taiwanese chipmaker has yet to reveal any details about the SoC, a recent leak sheds light on the performance gains it could offer over its current-generation chip, the Dimensity 9500. It is tipped to deliver peak clock speeds of up to 5GHz, achieving a multi core score between 12,000 and 12,500 points in benchmarks.

According to a Weibo post by tipster Digital Chat Station (translated from Chinese), the MediaTek Dimensity 9600 Pro could achieve single-core scores of around 4,200 to 4,300 points on Geekbench 6. Multi-core scores, meanwhile, are said to range between 12,000 and 12,500 points.

The tipster claims these figures are said to be based on early testing units and may not reflect the performance of the final retail unit. For comparison, the tipster noted that the previous generation chipset delivered roughly 4,000 points (single core) and 11,000 points (multi core).

On the other hand, the vanilla Oppo Find X9 scored 2,966 points (single core) and 8,710 points (multi core) in Gadgets 360’s tests. Meanwhile, the Find X9 Pro achieved 3,251 and 9,771 points in the same tests, respectively. The benchmark numbers hint towards performance at par, if not better, than the current generation Dimensity 9500.

The latest leak builds on earlier claims from the same source, which suggested that the Dimensity 9600 Pro could feature a peak clock speed of up to 5GHz. For comparison, the existing Dimensity 9500 features a prime core clocked at 4.21GHz, whereas the Snapdragon 8 Elite Gen 5 offers up to 4.74GHz clock speed.

MediaTek is reportedly using TSMC’s advanced N2P node to manufacture the flagship-grade chipset. It is tipped to have a 2+3+3 architecture with two ‘Canyon’ cores, three ‘Gelas-B cores and three ‘Gelas’ (translated from Chinese) cores, along with support for SME2 instructions, promising faster AI performance and ML efficiency.