Apple M3 Max

Apple M3 Max
Apple M3 Max processor review

Apple M3 Max: 16-Core CPU and 40-Core GPU

The Apple M3 Max was released in October 2023 as the highest-end processor in the M3 family. In its maximum configuration, it features a 16-core CPU with 12 performance cores and 4 efficiency cores, a 40-core GPU, 92 billion transistors, and support for up to 128 GB of unified memory with 400 GB/s of bandwidth.

This page examines the M3 Max version with 16 CPU cores and 40 GPU cores. The lower-end configuration with a 14-core CPU and 30-core GPU is covered separately.

Architecture and CPU

The main CPU change compared with the M2 Max is the increase in the number of performance cores from 8 to 12. The total core count increased from 12 to 16, while the number of efficiency cores remained unchanged at four.

The additional performance cores had a particularly strong effect on multi-threaded performance. In Cinebench R23, the MacBook Pro 16 with the M3 Max scores about 24,024 points, compared with 14,767 for the M2 Max model. This is an increase of approximately 63%. The Single-Core score rose from 1625 to 1968 points, an increase of about 21%.

Performance in Specific Devices

Device Processor Geekbench 6 SC Geekbench 6 MC Cinebench R23 SC Cinebench R23 MC Launch price
MacBook Pro 16 (2024) M4 Max 16 CPU / 40 GPU 4003 26 473 2209 27 595 $3999
MacBook Pro 16 (2023) M3 Max 16 CPU / 40 GPU 3096 21 243 1968 24 024 $3999
MacBook Pro 16 (2023) M2 Max 12 CPU / 38 GPU 2798 15 232 1625 14 767 $3499

In Geekbench 6, the M3 Max outperforms the M2 Max by approximately 11% in Single-Core and 39% in Multi-Core.

The 14-inch and 16-inch MacBook Pro models with the same 16-core M3 Max produce nearly identical results in short CPU tests. The difference appears under sustained full load: the larger MacBook Pro 16 chassis is better at maintaining maximum performance.

In Cinebench 2024, the MacBook Pro 16 scores about 1612 points in Multi-Core, while the MacBook Pro 14 scores about 1532. Single-Core results are virtually identical, at approximately 139-140 points.

The M4 Max later increased performance again. In Cinebench R23, its advantage over the M3 Max is approximately 12% in Single-Core and 15% in Multi-Core.

40-Core Graphics

The full version of the M3 Max features a 40-core GPU based on an architecture new to the Apple Silicon family. It introduced Dynamic Caching, hardware-accelerated ray tracing, and mesh shading support.

Graphics Tests

Device GPU Geekbench Metal Blender 3.3 Classroom Metal 3DMark Wild Life Extreme Unlimited
MacBook Pro 14 (2023) M3 Max 40-core GPU 100 376 29.6 s approximately 31 300
MacBook Pro 16 (2023) M3 Max 40-core GPU 99 672 32.0 s approximately 31 000
MacBook Pro 16 (2023) M2 Max 38-core GPU 86 678 51.2 s -

In Geekbench Metal, the M3 Max with its 40-core GPU outperforms the 38-core M2 Max by approximately 15%. In Blender, rendering time is reduced from about 51 seconds to 30-32 seconds.

The 14-inch and 16-inch MacBook Pro models with the same 40-core GPU produce similar results. There is no noticeable difference between the chassis in these short tests.

The GPU received a smaller speed increase than the CPU, but the M3 added hardware-accelerated ray tracing, Dynamic Caching, and mesh shading.

Memory

Memory bandwidth on the full M3 Max remained at the M2 Max level-up to 400 GB/s. At the same time, the maximum amount of unified memory increased from 96 to 128 GB.

The CPU and GPU share a common memory pool, allowing the graphics subsystem to work with large amounts of data without dedicated video memory. The 128 GB capacity is especially useful when rendering complex scenes and editing high-resolution video.

Conclusion

The CPU saw the largest improvement over the M2 Max. Increasing the number of performance cores from 8 to 12 raised the Cinebench R23 Multi-Core score by approximately 63%, while the Geekbench 6 Multi-Core score increased by about 39%.

The GPU received a more moderate speed increase: in Geekbench Metal, its advantage over the 38-core M2 Max is approximately 15%. At the same time, the M3 Max added Dynamic Caching, mesh shading, and hardware-accelerated ray tracing.

CPU Comparison

Compare Apple M3 Max with other CPUs

Please select CPU from the dropdown list.

Basic

Label Name
Apple
Platform
Laptop
Launch Date
October 2023
CPU Architecture
Apple M3 performance cores + efficiency cores
CPU Name
Apple M3 Max
Model Name
The Intel processor number is just one of several factors - along with processor brand, system configurations, and system-level benchmarks - to be considered when choosing the right processor for your computing needs.
Apple M3 Max
Foundry
TSMC
Generation
Apple M3 series

CPU Specifications

Total Cores
Cores is a hardware term that describes the number of independent central processing units in a single computing component (die or chip).
16
Total Threads
Where applicable, Intel® Hyper-Threading Technology is only available on Performance-cores.
16
Performance-cores
12
Efficient-cores
4
Efficient-core Max Turbo Frequency
Maximum E-core turbo frequency derived from Intel® Turbo Boost Technology.
2.75 GHz
Instruction Set Extensions
NEON
Performance-core Max Turbo Frequency
Maximum P-core turbo frequency derived from Intel® Turbo Boost Technology.
4.06 GHz
Extended Instruction Set
ARMv8-A, NEON
L1 Cache
P-cores: 192 KB instruction + 128 KB data per core; E-cores: 128 KB instruction + 64 KB data per core
L2 Cache
P-core cluster: 36 MB; E-core cluster: 4 MB
L3 Cache
48 MB system level cache
Technology
Lithography refers to the semiconductor technology used to manufacture an integrated circuit, and is reported in nanometer (nm), indicative of the size of features built on the semiconductor.
3 nm
Instruction Set
The instruction set is a hard program stored inside the CPU that guides and optimizes CPU operations. With these instruction sets, the CPU can run more efficiently. There are many manufacturers that design CPUs, which results in different instruction sets, such as the 8086 instruction set for the Intel camp and the RISC instruction set for the ARM camp. x86, ARM v8, and MIPS are all codes for instruction sets. Instruction sets can be extended; for example, x86 added 64-bit support to create x86-64. Manufacturers developing CPUs that are compatible with a certain instruction set need authorization from the instruction set patent holder. A typical example is Intel authorizing AMD, enabling the latter to develop CPUs compatible with the x86 instruction set.
ARMv8-A
Transistor Count
92 billion

Memory Specifications

Memory Bus Width
512-bit
Memory Type
Intel® processors come in four different types: Single Channel, Dual Channel, Triple Channel, and Flex Mode. Maximum supported memory speed may be lower when populating multiple DIMMs per channel on products that support multiple memory channels.
Unified LPDDR5-6400
LPDDR5 Speed
LPDDR5-6400
Max Memory Size
Max memory size refers to the maximum memory capacity supported by the processor.
128 GB
Max Memory Bandwidth
Max Memory bandwidth is the maximum rate at which data can be read from or stored into a semiconductor memory by the processor (in GB/s).
400 GB/s
Maximum Memory Speed
6400 MT/s

GPU Specifications

External Display Standard
Thunderbolt 4, HDMI 2.1
GPU Name
Apple M3 Max GPU
Max External Display Resolution
Up to three 6K 60Hz displays plus one 4K 144Hz display; or up to two 6K 60Hz displays plus one 8K 60Hz / 4K 240Hz display
Video Concurrency
Multiple streams of 4K and 8K ProRes video; AV1 decode
Video Decode
H.264, HEVC, ProRes, ProRes RAW, AV1 decode
Video Encode
H.264, HEVC, ProRes, ProRes RAW
Video Processing Unit
Apple media engine with ProRes acceleration and AV1 decode
GPU Max Dynamic Frequency
1400 MHz
Graphics Base Frequency
Graphics Base frequency refers to the rated/guaranteed graphics render clock frequency in MHz.
390 MHz
Graphics Core Count
40
Number of Displays Supported
Up to 4 external displays
GPU APIs
Metal, OpenCL
Graphics Performance
Up to 14.2 TFLOPS FP32
Media Engine
Hardware-accelerated H.264, HEVC, ProRes, ProRes RAW, and AV1 decode
Video Decode Engines
1
Video Encode Engines
2
ProRes Encode/Decode Engines
2
OpenCL Support
OpenCL (Open Computing Language) is a multi-platform API (Application Programming Interface) for heterogeneous parallel programming.
Yes

AI Specifications

AI Engine
16-core Apple Neural Engine
Neural Engine Core Count
16
NPU Name
Apple Neural Engine
NPU Performance
18 TOPS

Connectivity

Bluetooth Support
Yes
Bluetooth Version
Bluetooth 5.3
Wi-Fi Standard
Wi-Fi 6E (802.11ax)

Interfaces and Ports

Thunderbolt Support
Yes, Thunderbolt 4 up to 40 Gb/s
USB Version
USB4
USB4 Support
Yes, USB4 up to 40 Gb/s

Miscellaneous

Hardware-Verified Secure Boot
Yes, Apple silicon secure boot chain of trust
Runtime Anti-Exploitation
Kernel Integrity Protection, Pointer Authentication Codes, Fast Permission Restrictions

Benchmarks

Cinebench R23
Single Core Score
1968
Cinebench R23
Multi Core Score
24028
Geekbench 6
Single Core Score
3128
Geekbench 6
Multi Core Score
20969
Geekbench 5
Single Core Score
2150
Geekbench 5
Multi Core Score
22736
Passmark CPU
Single Core Score
4784
Passmark CPU
Multi Core Score
41257
Cinebench 2024
Single Core Score
141
Cinebench 2024
Multi Core Score
1607
Cinebench 2024 GPU
Score
12980
Blender
Score
420

Compared to Other CPU

Cinebench R23 Single Core
2634 +33.8%
2042 +3.8%
1968
1674 -14.9%
1373 -30.2%
Cinebench R23 Multi Core
45651 +90%
24028
16068 -33.1%
13316 -44.6%
M3
10437 -56.6%
Geekbench 6 Single Core
4698 +50.2%
3219 +2.9%
3128
Geekbench 6 Multi Core
38432 +83.3%
20969
16988 -19%
15550 -25.8%
Geekbench 5 Single Core
2974 +38.3%
2150
1875 -12.8%
1793 -16.6%
1705 -20.7%
Geekbench 5 Multi Core
55920 +146%
31330 +37.8%
22736
19031 -16.3%
Passmark CPU Single Core
6341 +32.5%
4784
4598 -3.9%
4453 -6.9%
4287 -10.4%
Passmark CPU Multi Core
47554 +15.3%
44719 +8.4%
41257
38100 -7.7%
35626 -13.6%
Cinebench 2024 Single Core
142 +0.7%
141
M3
137 -2.8%
134 -5%
Cinebench 2024 Multi Core
1756 +9.3%
1607
1431 -11%
1335 -16.9%
Cinebench 2024 GPU
22575 +73.9%
12980
12948 -0.2%
8668 -33.2%
7768 -40.2%
Blender
1154 +174.8%
420
201 -52.1%
151 -64%
71 -83.1%