Apple M4 (8 cores) vs Qualcomm Snapdragon 8 Elite for Galaxy vs Apple M4 (9 cores)
Apple M4 (8 cores)
► remove from comparison
The 8-core Apple M4 is an ARM architecture processor (SoC) that sports 8 CPU cores along with an 8-core GPU with hardware RT support and other modern features. A 16-core neural engine, USB 4 and Thunderbolt 4 are onboard, too.
The processor debuted in October 2024; it cuts down on the number of efficient cores but thankfully has the same 4 performance cores that its more powerful M4 brothers (non-Pro, non-Max) have. The P-cores probably run at 4.1 GHz or more; we don't have exact figures as of yet. The E-cores probably run at sub-3 GHz clock speeds.
This is the third member of the M4 (non-Pro, non-Max) family. It joins the 10-core M4 and the 9-core M4 that were unveiled several months before it.
The chip is believed to be based on the ARM v9.4-A microarchitecture to a certain extent. It comes with 16 GB or 24 GB of fast on-package LPDDR5x-7500 RAM; M3 processors had to be content with 6400 MT/s, for reference, while higher end M4 Pro and M4 Max processors get LPDDR5x-8533. It is not yet clear if this new chip is a 10-core M4 with several modules disabled, or if it's actually a new die that has fewer CPU and GPU cores by design.
Its performance is set to be very close to what Intel Lunar Lake processors such as the 256V deliver.
As far as power consumption is concerned, the SoC probably eats about 10 W when under long-term workloads. We'll update this section once we have one of the new Macs in for testing.
Qualcomm Snapdragon 8 Elite for Galaxy
► remove from comparison
The Snapdragon 8 Elite for Galaxy is a high-end SoC for Samsung's flagship smartphones and tablets of 2025.
For the first time, Qualcomm is using CPU cores that are not adapted ARM standard designs, but have their own ARM-based architecture: The manufacturer calls the processor part second-generation Oryon CPU. The previously used efficiency cores are no longer required and their task is taken over by the performance cores, which, according to the manufacturer, are now flexible and efficient enough for this purpose. The differences to the conventional Snapdragon 8 Elite are reflected in a slightly higher performance. The prime cores, GPU and NPU clock slightly higher.
There are a total of 2 prime cores with up to 4.47 GHz, which according to Qualcomm can also be overclocked at times. 6 performance cores for less demanding tasks are also installed, clocking at up to 3.53 GHz. Both processor clusters have 12 MB L2 cache, which should offer low latency despite its size. In the next stage, an 8 MB SLC L3 cache awaits the data before it has to go to the main memory via four 16-bit LPDDR5x memory controllers.
The GPU is the Adreno 830 GPU is used, which has also been redesigned. Thanks to the so-called "sliced architecture", it now offers three computing cores with a clock speed of up to 1.2 GHz.
The focus for the 2025 flagship smartphones is also on AI. The new Hexagon NPU is said to offer 45% more performance compared to its predecessor and also work 45% more efficiently. There is also broad support for multimodal generative AI models.
The three ISPs have also been revised and are now called Spectra AI ISP. They can handle data from up to 3 48-megapixel cameras simultaneously. A resolution of up to 320 megapixels is possible with a single camera. Unlimited semantic segmentation is possible with 4K images, meaning that as many different layers and objects can be distinguished as there are in the image. This allows automatic photo optimization to be improved. In addition, the NPU now has direct access to the RAW data from the sensors for the first time.
The integrated Snapdragon X80 Modem-RF 5G system allows simultaneous transfers of 5G-mmWave and 5G-Sub6, enabling data rates of up to 10 GBit/s in download and up to 3.5 GBit/s in upload. The modem is also able to support all globally known 5G bands. Qualcomm's FastConnect 7900, which supports Bluetooth 6.0 and Wi-Fi 7, is responsible for local communication.
The SoC is manufactured at TSMC using the 3 nm process (N3E).
Apple M4 (9 cores)
► remove from comparison
The 9-core Apple M4 is a rather fast ARM architecture processor (SoC) that sports 9 CPU cores, a 16-core neural engine and a 10-core GPU with hardware RT support and other modern features. Wi-Fi 6E and Bluetooth 5.3 as well as Thunderbolt 3 and USB 4 are all onboard, too. The M4 debuted in May 2024 as part of an iPad launch event; it has 3 performance cores running at a clock speed of up to 4.3 GHz and 6 efficient cores running at way under 3 GHz whereas the M3 (10 GPU cores) has 4 performance cores and 4 efficiency cores.
The faster 10-core M4 chip delivers 10% higher multi-thread performance and 15% higher graphics performance, to give you a rough figure. More importantly, the 10-core M4 can be had in systems with an active cooling solution, such as Fall 2024 MacBook Pros, whereas the 9-core one is an iPad Pro exclusive. Systems with a fan deliver much higher sustained performance than the ones without it.
Architecture and Features
The new CPU cores run at faster clock speeds than what the M3 was capable of and they also feature minor architectural improvements. The CPU cores are thought to be based on the ARM v9.4-A microarchitecture to a certain degree. The M4 comes with 16 GB or more (depending on the SKU) of on-package LPDDR5x-7500 RAM whereas the M3 was limited to 6400 MT/s. The updated NPU delivers up to 38 TOPS of performance for AI workloads.
Performance
Single-thread performance, multi-thread performance and NPU performance all got a noticeable boost compared to what we had with the M3 (10 GPU cores). The new processor is about 11% faster than the M3 in short-term multi-thread workloads while besting every M3 series chip possible in single-thread tasks by a comfortable margin. Geekbench 6.2 Multi puts the 9-core M4 right above the Core i9-13900H and the Ryzen 7 7840HS; in fact, the M4 is just 3% slower than Intel's top-of-the-line Core Ultra 9 185H chip. A 13% to 18% single-thread performance improvement over M3 series chips is evident if we look at Mozilla's Kraken test results. Octane V2 seems to think the M4 is just 4% slower than Intel's mighty Core i9-14900HX. CrossMark results suggest the M4 is about as fast as the Core i9-13900H.
The 10-core M4 is about 10% faster in multi-threaded tasks than the 9-core part is. It also delivers just slightly higher single-thread performance.
While Apple undoubtedly deserves some praise for what it managed to do here, it is important to highlight that all of the tests we did involve short-term workloads only. The M4 will suffer from heavy throttling if subjected to long-term workloads as there is no active cooling inside that super-thin iPad Pro case.
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".
As far as performance is concerned, it appears the 9-core M4's GPU runs at significantly lower clock speeds than the (otherwise the same) GPU built into the 10-core M4 chip. This leads to a disappointing situation where what's supposedly a newer graphics adapter is about 10% slower than the 10-core GPU built into the M3. Still, this M4 GPU delivers very decent benchmark scores that put it in the same ballpark as the Radeon RX 6500M.
Power consumption
The chip's sustained power consumption is limited to ~7 W, with short-term peaks of up to 14 W possible.
The chip is built with a "second generation" 3 nm TSMC process that's still cutting-edge as of late 2024.
Model | Apple M4 (8 cores) | Qualcomm Snapdragon 8 Elite for Galaxy | Apple M4 (9 cores) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Series | Apple M4 | Qualcomm Snapdragon 8 | Apple M4 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Series: M4 |
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Clock | 2900 - 4000 MHz | 3530 - 4470 MHz | 2900 - 4400 MHz | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L2 Cache | 4 MB | 12 MB | 4 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Cores / Threads | 8 / 8 4 x 4.4 GHz Apple M4 P-Core 2.9 GHz Apple M4 E-Core | 8 / 8 3 x 2 x 4.5 GHz Qualcomm Oryon Gen 2 Prime 6 x 3.5 GHz Qualcomm Oryon Gen 2 Performance | 9 / 9 3 x 4.4 GHz Apple M4 P-Core 6 x 2.9 GHz Apple M4 E-Core | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TDP | 5 Watt | 9 Watt | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
TDP Turbo PL2 | 15 Watt | 14 Watt | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Technology | 3 nm | 3 nm | 3 nm | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Features | Unified Memory LPDDR5X-7500 (120 GB/s), 16-Core Neural Engine, Media Engine (Encoding / Decoding: H.264, HEVC, ProRes, ProRes RAW, AV1 Decoding only) | Adreno GPU, Spectra ISP, Hexagon, X80 5G Modem, Wi-Fi 7, LPDDR5x 4800 | Unified Memory LPDDR5X-7500 (120 GB/s), 16-Core Neural Engine, Media Engine (Encoding / Decoding: H.264, HEVC, ProRes, ProRes RAW, AV1 Decoding only) | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
iGPU | Apple M4 8-core GPU | Qualcomm Adreno 830 ( - 1100 MHz) | Apple M4 10-core GPU | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Architecture | ARM | ARM | ARM | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Announced | |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Codename | Oryon Gen 2 | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
L3 Cache | 8 MB | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Manufacturer | www.qualcomm.com |
Benchmarks
Average Benchmarks Qualcomm Snapdragon 8 Elite for Galaxy → 0% n=0
Average Benchmarks Apple M4 (9 cores) → 0% n=0

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