Apple M2 Pro, M2 Max chips announced for next generation of Macs

Apple has taken the lid off new custom silicone chips for new MacBook Pro models and a Mac mini. Apple’s M2 Pro and M2 Max feature a more powerful CPU and GPU and up to 96GB of unified memory.

The new Apple M2 Pro chip scales up the architecture of M2, delivering up to 12-core CPU and up to 19-core GPU with up to 32GB of fast unified memory. M2 Max, on the other hand, builds on the capabilities of M2 Pro and adds up to 38-core GPU, and supports up to 96GB of fast unified memory.

The M2 Pro chip consists of 40 billion transistors, which is nearly 20 percent more than that on the M1 Pro. It features 200GB/s of unified memory bandwidth, twice that of M2, and up to 32GB of low-latency unified memory. The 12-core CPU consists of up to eight high-performance cores and four high-efficiency cores.

This makes it 20 percent faster than the 10-core CPU on the M1 Pro in multithreaded CPU performance. The GPU in M2 Pro can be configured with up to 19 cores and includes a larger L2 cache. Apple claims graphics speeds are up to 30 percent faster than that of the M1 Pro.

The Apple M2 Max has 67 billion transistors, which is 10 billion more than the M1 Max. Its 400GB/s of unified memory bandwidth is twice that of M2 Pro and supports up to 96GB of fast unified memory. Additionally, the M2 Max features a 12-core CPU paired with up to a 38-core GPU and a larger L2 cache. Apple claims that the new GPU delivers 30 percent faster graphics speeds than the M1 Max. Apple also says that the M2 Max is the “world’s most powerful and efficient chip” for a pro laptop.

The M2 Pro and M2 Max chips feature Apple’s next-gen 16-core Neural Engine that is up to 40 percent faster than the previous generation. M2 Pro has an immensely powerful and efficient media engine. Moreover, the M2 Max features two videos encode engines and two ProRes engines, bringing up to 2x faster video encoding than M2 Pro. The new M2 Pro and M2 Max chips will be found on Apple’s new MacBook Pro and Mac mini models.