M2 Max

Announced January 2023. Follows the M1 Max.

  • 12-core

    CPU

  • 38-core

    GPU

  • 96GB

    Max memory

  • 5 nm

    Process

Performance

  • 2,699

    Single-core

    11% faster than M1 Max

  • 15,184

    Multi-core

    17% faster than M1 Max

  • 151,394

    Graphics

    25% faster than M1 Max

Geekbench 6 scores, measured on Mac Studio (2023).

CPU

12 cores

8 performance, 4 efficiency

Up to 3.69 GHz

Graphics

38 cores

Apple-designed

Neural Engine

16 cores

15.8 trillion operations per second

Memory

Up to 96GB

LPDDR5-6400

400GB/s bandwidth

Silicon

5 nm

TSMC N5P

67 billion transistors

Ships as

  • 12-core CPU · 38-core GPU · Up to 96GB

    The full chip

  • 12-core CPU · 30-core GPU · Up to 64GB

Found in

  • MacBook Pro 14-inch (2023)
  • MacBook Pro 16-inch (2023)
  • Mac Studio (2023)

The M Max line

The same generation

CPU performance over the years

Geekbench 6 multi-core score, from each chip’s best-cooled device. No Geekbench 6 score exists for chips before the A8 (2014), or for Watch chips.

Show the numbers
Geekbench 6 scores
ChipYearMulti-coreDevice
A82014614iPad mini 4
A8X2014963iPad Air 2
A92015990iPad (5th generation)
A9X20151,314iPad Pro 12.9-inch (1st generation)
A10 Fusion20161,424iPad (7th generation)
A10X Fusion20172,358iPad Pro 12.9-inch (2nd generation)
A11 Bionic20172,153iPhone X
A12 Bionic20182,914iPhone XS Max
A12X Bionic20184,652iPad Pro 12.9-inch (3rd generation)
A13 Bionic20194,128iPhone 11 Pro Max
A12Z Bionic20204,914iPad Pro 12.9-inch (4th generation)
A14 Bionic20205,123iPhone 12 Pro Max
M120208,650Mac mini (2020)
A15 Bionic20215,941iPhone 13 Pro Max
M1 Max202112,923Mac Studio (2022)
M1 Pro202112,393MacBook Pro 16-inch (2021)
M1 Ultra202218,870Mac Studio (2022)
M220229,964Mac mini (2023)
A16 Bionic20226,926iPhone 14 Pro Max
M2 Max202315,184Mac Studio (2023)
M2 Pro202314,879Mac mini (2023)
M2 Ultra202322,289Mac Studio (2023)
A17 Pro20237,407iPhone 15 Pro Max
M3 Max202320,774MacBook Pro 16-inch (Nov 2023)
M3 Pro202315,646MacBook Pro 16-inch (Nov 2023)
M3202311,784MacBook Pro 14-inch (Nov 2023)
M4202415,099Mac mini (2024)
A18 Pro20248,807iPhone 16 Pro Max
A1820248,441iPhone 16 Plus
M4 Max202426,669Mac Studio (2025)
M4 Pro202422,960Mac mini (2024)
M3 Ultra202529,449Mac Studio (2025)
A19 Pro202510,034iPhone 17 Pro Max
A1920259,406iPhone 17
M5202517,820MacBook Pro 14-inch (2025)
M5 Max202629,690Mac Studio (2026)
M5 Pro202628,900Mac mini (2026)
M5 Ultra202638,144Mac Studio (2026)
M6202620,970Mac mini (2026)
A20 Pro202612,789iPhone 18 Pro Max

Graphics performance vs. GPU cores

Geekbench 6 Metal score. One line per Mac generation, from its base chip to its Ultra; iPhone and iPad chips are the dots. A8–A10X Fusion are left out: their GPUs were licensed PowerVR designs, counted in clusters rather than cores.

Show the numbers
Metal scores and GPU cores
ChipGPU coresMetalPer core
A11 Bionic35,2241,741
A12 Bionic48,6832,171
A12X Bionic717,4932,499
A13 Bionic413,7333,433
A12Z Bionic819,2862,411
A14 Bionic417,0464,262
M1834,5204,315
A15 Bionic521,9114,382
M1 Pro1671,1094,444
M1 Max32121,3683,793
M1 Ultra64180,6192,822
M21048,6744,867
A16 Bionic524,3484,870
M2 Pro1986,4124,548
M2 Max38151,3943,984
M2 Ultra76227,1242,988
A17 Pro628,6364,773
M31050,1135,011
M3 Pro1882,2724,571
M3 Max40164,0904,102
M41058,3255,833
A18528,8255,765
A18 Pro633,1535,526
M4 Pro20112,2145,611
M4 Max40198,3894,960
M3 Ultra80260,0033,250
A19538,3137,663
A19 Pro647,2437,874
M51078,1787,818
M5 Pro20139,3906,970
M5 Max40234,3005,858
M5 Ultra80335,0264,188
M61298,1368,178
A20 Pro763,7039,100

Transistors, to scale

Each square’s area is the chip’s transistor count, in billions; the squares are not die sizes. 37 chips have no published count, among them every chip since the M3 Ultra (2025).

Show the numbers
Transistor counts
ChipYearTransistors
A720131B
A820142B
A8X20143B
A10 Fusion20163.3B
A11 Bionic20174.3B
A12 Bionic20186.9B
A12X Bionic201810B
A13 Bionic20198.5B
A12Z Bionic202010B
A14 Bionic202011.8B
M1202016B
A15 Bionic202115B
M1 Pro202133.7B
M1 Max202157B
M1 Ultra2022114B
M2202220B
A16 Bionic202216B
M2 Pro202340B
M2 Max202367B
M2 Ultra2023134B
A17 Pro202319B
S920235.6B
M3202325B
M3 Pro202337B
M3 Max202392B
M4202428B
M3 Ultra2025184B

Process node over the years

The smallest node any Apple chip was in production on, on a logarithmic scale. A picked chip under the line was made on an older node.

Show the numbers
Process nodes
ChipYearNode
A4201045nm (45 nm)
A5201145nm (45 nm)
A5X201245nm (45 nm)
A6201232nm HKMG (32 nm)
A6X201232nm HKMG (32 nm)
A7201328nm HKMG (28 nm)
A8201420SoC (20 nm)
S1201428nm HKMG (28 nm)
A8X201420SoC (20 nm)
A9201514LPE / 16FF+ (14 nm)
A9X201516FF+ (16 nm)
A10 Fusion201616FFC (16 nm)
A10X Fusion201710FF (10 nm)
A11 Bionic201710FF (10 nm)
A12 Bionic2018N7 (7 nm)
S42018N7 (7 nm)
A12X Bionic2018N7 (7 nm)
A13 Bionic2019N7P (7 nm)
S52019N7 (7 nm)
A12Z Bionic2020N7 (7 nm)
A14 Bionic2020N5 (5 nm)
S62020N7P (7 nm)
M12020N5 (5 nm)
A15 Bionic2021N5P (5 nm)
S72021N7P (7 nm)
M1 Pro2021N5 (5 nm)
M1 Max2021N5 (5 nm)
M1 Ultra2022N5 (5 nm)
M22022N5P (5 nm)
A16 Bionic2022N4P (4 nm)
S82022N7P (7 nm)
M2 Pro2023N5P (5 nm)
M2 Max2023N5P (5 nm)
M2 Ultra2023N5P (5 nm)
A17 Pro2023N3B (3 nm)
S92023N4 (4 nm)
M32023N3B (3 nm)
M3 Pro2023N3B (3 nm)
M3 Max2023N3B (3 nm)
M42024N3E (3 nm)
A182024N3E (3 nm)
A18 Pro2024N3E (3 nm)
S102024N4 (4 nm)
M4 Pro2024N3E (3 nm)
M4 Max2024N3E (3 nm)
C12025N4 (4 nm)
M3 Ultra2025N3B (3 nm)
A192025N3P (3 nm)
A19 Pro2025N3P (3 nm)
M52025N3P (3 nm)
M5 Pro2026N3P (3 nm)
M5 Max2026N3P (3 nm)
M5 Ultra2026N3P (3 nm)
M62026N2 (2 nm)
A20 Pro2026N2 (2 nm)

Sources

The chip drawings under CPU and Graphics are illustrations: the core counts are the chip's own, the layout and size are not the real die's.