Apple is developing an A18 chip for the iPhone 16 series models launching this year. The chip is rumored to have entered the testing phase early, showing better single-core performance scores than the M3 series in Geekbench 6 benchmarks, although its multi-core performance scores are less impressive. Reports suggest only a 10% increase in multi-core performance compared to the A17 Pro.

According to Wccftech, the A18 chip's neural network engine is exceptionally powerful, with arithmetic capabilities that may surpass the M4 chip. The neural network engine is crucial for handling generative AI functions, meaning the iPhone 16 series models will be able to run AI models locally, adapting better to various types of AI tasks.

Apple introduced the M4 chip in the new iPad Pro at its "Let Loose" special event in May. Built on a second-generation 3nm process with 28 billion transistors, it utilizes an SoC architecture to further improve power efficiency and incorporates Apple's fastest neural network engine to date, operating at up to 38 trillion operations per second. If the A18 chip features an even more powerful neural network engine, its computing speed will be even higher.

Recent news indicates that Apple might take a different approach in marketing the A18 chip, stepping away from the usual "powerful" or "very efficient" descriptions typically used for new SoCs. Instead, Apple may focus on highlighting the improved performance of the neural network engine, potentially opting for a larger chip design.