Apple M2 Max vs Apple M4 (10 cores) vs Apple M4 Max (16 cores)
Apple M2 Max
► remove from comparison
The Apple M2 Max is a System on a Chip (SoC) from Apple that is found in the early 2023 MacBook Pro 14 and 16-inch models. It offers all 12 CPU cores available in the chip divided in eight performance cores (P-cores) and four power-efficiency cores (E-cores). The E-cores clock with up to 3.4 GHz, the P-Cores up to 3.7 GHz (mostly 3.3 GHz in multi-threaded workloads and 3.4 GHz in single threaded).
The big cores (codename Avalanche) offer 192 KB instruction cache, 128 KB data cache, and 36 MB shared L2 cache (up from 24 MB in the M1 Pro). The four efficiency cores (codename Blizzard) are a lot smaller and offer only 128 KB instruction cache, 64 KB data cache, and 4 MB shared cache. CPU and GPU can both use the 49 MB SLC (System Level Cache).
The unified memory (32, 64, or 96 GB LPDDR5-6400) next to the chip is connected by a 512 Bit memory controller (400 GB/s bandwidth) and can be used by the GPU and CPU.
The CPU performance should be quite similar to the M2 Pro as only the higher memory bandwidth and bigger L3 cache could make a difference for some workloads.
The integrated graphics card in the M1 Max offers a 30-core GPU or a GPU with all 38 cores.
Furthermore, the SoC integrates a fast 16 core neural engine (faster than M1 Max), a secure enclave (e.g., for encryption), a unified memory architecture, Thunderbolt 4 controller, an ISP, and media de- and encoders (including ProRes).
The M2 Max is manufactured in 5 nm at TSMC (second generation) and integrates 40 billion transistors. The power consumption of the CPU part is up to 36 Watt according to powermetrics. When fully loading the CPU and GPU cores, the chip uses up to 89 Watt and the CPU part is limited to 25 Watt.
Apple M4 (10 cores)
► remove from comparison
The 10-core Apple M4 is an impressively fast ARM architecture processor (SoC) sporting 4 performance and 6 efficient CPU cores along with a 16-core neural engine and a 10-core GPU with hardware RT support and other modern features. On-package 7500 MT/s LPDDR5x RAM, USB 4 support and Thunderbolt 4 support are onboard as well, as are Wi-Fi 6E and Bluetooth 5.3.
The M4 debuted in May 2024 as part of an iPad launch event; its 4 performance cores run at up to 4.4 GHz, representing a healthy improvement over the M3 (10-core GPU) and its 4.06 GHz top clock speed. Besides, this new chip also has 6 efficient cores at its disposal that run at up to 2.9 GHz. The M3 had to make do with 4 performance cores and 4 efficient cores, for reference.
Architecture and Features
The new CPU cores run at faster clock speeds than what the M3 was capable of while also featuring some minor architectural improvements. A heavily customized version of ARM's v9.4-A microarchitecture is reportedly employed here. The M4 comes with on-package LPDDR5x-7500 RAM (120 GB/s) whereas the M3 was limited to 6400 MT/s. The updated NPU delivers up to 38 TOPS of performance for AI workloads.
Performance
In 2024 iPad Pros, the 3 nm chip is 10% to 20% faster than the M3 (10 GPU cores) in both multi-thread and single-thread short-term workloads. It also delivers about 10% higher multi-thread performance than the 9-core M4. This puts the 10-core part on an equal footing with the Ryzen 7 7745HX, Core i7-13705H and other higher-end x86 processors released in 2023 and 2024.
iPad Pros come with no active cooling solution however Mac minis and MacBook Pros do have one and the latest iMacs probably do as well. As our Fall 2024 Mac mini review shows, the M4's short-term performance doesn't benefit much from the presence of a fan, however its sustained performance benefits tremendously.
Graphics
The M4 GPU (10 cores) has hardware support for ray tracing as well as mesh shading and other modern technologies. It supports external displays with resolutions as high as 8K.
The graphics adapter runs at higher clock speeds than what the (otherwise identical) 10-core GPU built into the 9-core M4 can muster. As our Mac mini 2024 review confirmed, its performance is good enough to run many 2023 and 2024 games at 1080p on low or medium quality settings.
- Сyberpunk 2077 2.1 Phantom Liberty (1080p, Med): 28 fps
Much like it is with CPU performance, M4-powered iPad Pros will be much slower in long-term GPU-intensive workloads than iMacs, Mac minis and MacBook Pros due to the lack of proper cooling.
Power consumption
This specific M4 chip is found in 2024 iPad Pros, 2024 iMacs, 2024 Mac minis and 2024 MacBook Pros. Within the thin metal cases of the iPads - no active cooling - the chip's sustained power consumption hovers in the 4 W - 5 W area, with short-term peaks of up to 14 W possible. However, Mac minis and MacBook Pros do have an active cooling solution (and the new iMacs probably do, too) meaning the power consumption figures are several times higher. In our Fall 2024 Mac mini review, the SoC consumed 30 W to 40 W when under high loads.
The M4 is built with a "second generation" 3 nm TSMC process which is still cutting-edge as of late 2024.
Apple M4 Max (16 cores)
► remove from comparisonThe 16-core Apple M4 Max is a powerful ARM architecture processor (SoC) for laptops and mini-PCs that debuted in Sep 2024. It features 12 performance CPU cores running at well over 4 GHz along with 4 efficient cores running at under 3 GHz. The 40-core M4 Max GPU and at least 48 GB of fast 546 GB/s LPDDR5x on-package memory depending on the configuration are included as well, as is USB 4 and Thunderbolt 5 support.
The built-in 16 core neural engine (up to 38 TOPS) is found across the whole M4 chip family. Furthermore, all M4 processors are thought to be based on the ARM v9.4-A architecture to a certain degree.
Performance
The M4 Max is set to be just around 15% faster than the 16-core M3 Max. This is very, very impressive.
Graphics
The 40-core M4 Max GPU has hardware support for ray tracing as well as mesh shading and other modern technologies. It supports external displays with resolutions as high as "8K" and it can HW-decode a few popular video codecs such as h.264, h.265 and AV1.
Its gaming performance is set to be about as good the GeForce RTX 4070 Laptop. Please keep in mind that very few games have been compiled specifically for Apple silicon Macs meaning most titles have to be run via emulation layers. Some of them are displayed with visual artefacts as a result, or don't start at all.
Power consumption
To let so many CPU and GPU cores breathe, some really high TDP figures are required. We assume the chip consumes 90 W to 100 W when under full load.
The 2nd generation 3 nm TSMC process the M4 Max is built with delivers good power efficiency, as of late 2024.
Model | Apple M2 Max | Apple M4 (10 cores) | Apple M4 Max (16 cores) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Series | Apple M2 | Apple M4 | Apple M4 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Series: M4 |
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Clock | 2424 - 3696 MHz | 2890 - 4464 MHz | 2592 - 4512 MHz | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L1 Cache | 3.3 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L2 Cache | 36 MB | 4 MB | 4 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L3 Cache | 48 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cores / Threads | 12 / 12 | 10 / 10 4 x 4.5 GHz Apple M4 P-Core 6 x 2.9 GHz Apple M4 E-Core | 16 / 16 12 x 4.5 GHz Apple M4 P-Core 4 x 2.6 GHz Apple M4 E-Core | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TDP | 79 Watt | 40 Watt | 90 Watt | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Transistors | 67000 Million | 28 Million | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Technology | 5 nm | 3 nm | 3 nm | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Features | ARMv8 Instruction Set | Unified Memory LPDDR5X-7500 (120 GB/s), 16-Core Neural Engine, Media Engine (Encoding / Decoding: H.264, HEVC, ProRes, ProRes RAW, AV1 Decoding only) | Unified Memory LPDDR5X-8533 (546 GB/s), 16-Core Neural Engine, Dual Media Engine (Encoding / Decoding: H.264, HEVC, ProRes, ProRes RAW, AV1 Decoding only) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
iGPU | Apple M2 Max 38-Core GPU | Apple M4 10-core GPU | Apple M4 Max 40-Core GPU | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Architecture | ARM | ARM | ARM | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Announced | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TDP Turbo PL2 | 40 Watt |
Benchmarks
Average Benchmarks Apple M2 Max → 100% n=14
Average Benchmarks Apple M4 (10 cores) → 118% n=14
Average Benchmarks Apple M4 Max (16 cores) → 154% n=14

* Smaller numbers mean a higher performance
1 This benchmark is not used for the average calculation