Quick/Comparisons
50 | Gaming | 28 |
59 | Workstation | 30 |
41 | AI/ML |
Compare with something else?
Compare Apple M3 Ultra with another →27 |
61 | Energy Efficiency | 56 |
52 | Quick Comparison Final Score | 33 |
| Custom | Graphics Processor | AD107 (GN21-X2) |
| 10240 | Cores | 2560 |
| 640 / 320 | TMUs / ROPs | 80 / 32 |
| 7360.18 | Blender GPU | 2575.82 |
| — | Forza Horizon 5 | 66 |
| — | The Witcher 3 | 92 |
| — | Counter-Strike 2 | 125 |
| — | Far Cry 6 | 79 |
| — | Hogwarts Legacy | 51 |
| — | Call of Duty | 64 |
| — | Ghost of Tsushima | 43 |
| — | Cyberpunk 2077 | 56 |
| — | Tomb Raider | 93 |
| — | Average FPS | 74 |
| — | 1080p High | 93 |
| — | 1080p Ultra | 74 |
| — | 1440p Ultra | 54 |
| — | 4K Ultra | 30 |
| — | Margin of Error | Low |
| 147719 | GB6 Compute Score | 81436 |
| 204.5 img/sec | Background Blur | 141.4 img/sec |
| 126 img/sec | Face Detection | 84.1 img/sec |
| 6.53 Gpixels/sec | Horizon Detection | 3.07 Gpixels/sec |
| 10.7 Gpixels/sec | Edge Detection | 3.91 Gpixels/sec |
| 9.5 Gpixels/sec | Gaussian Blur | 3.85 Gpixels/sec |
| 1.05 Gpixels/sec | Feature Matching | 1.1 Gpixels/sec |
| 560.3 Gpixels/sec | Stereo Matching | 325.9 Gpixels/sec |
| 25267.8 FPS | Particle Physics | 11008.8 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 14319 |
| — | G2D Mark | 630 |
| — | DirectX 11 | 129 FPS |
| — | DirectX 12 | 61 FPS |
| — | GPU Compute | 5914 Ops/s |
| 21356 | GB6 ML Single Precision | 15570 |
| 23530 | GB6 ML Half Precision | 27160 |
| 21706 | GB6 ML Quantized | 12230 |
| 8094 | Image Classification (SP) | 6233 |
| 24595 | Image Segmentation (HP) | 16395 |
| 25190 | Image Super Resolution (Q) | 18240 |
| 38337 | Face Detection (HP) | 30758 |
| 111021 | Pose Estimation (Q) | 61300 |
| 2786 | Text Classification (SP) | 3124 |
| 4339 | Machine Translation (HP) | 4357 |
| 7668 | Object Detection (SP) | 7812 |
| 37488 | Depth Estimation (Q) | 29192 |
| 244483 | Style Transfer (SP) | 117759 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 442 GPixel/s | Pixel Fill Rate | 51-76 GPixel/s |
| 883 GTexel/s | Texture Fill Rate | 128-190 GTexel/s |
| 28.3 TFLOPS | FLOPS (FP32) | 8.2-12.1 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 6 GB |
| 6400 MHz | Memory Clock | 2000 MHz |
| — | Effective Memory Speed | 16000 Mbps |
| 1024-bit | Bus | 96-bit |
| No | ECC | No |
| 819.3 GB/s | Memory Bandwidth | 192 GB/s |
| Custom | Interface | PCIe 4.0 x16 |
| 140 W | TGP | 35-140 W (configurable) |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 5 nm |
| — | Die Size | 156 mm² |
| 184 billion | Transistor Count | 18.9 billion |
| — | Transistor Density | 121.15 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | 8.9 |
| Yes | Ray Tracing | Yes |
| No | DLSS | DLSS 3 |
| — | DisplayPort | 1.4a |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| March 5, 2025 | Released | February 22, 2023 |
| Desktop | Case | Laptop |
| Professional | Purpose | Gaming |
| High-end | Segment | Entry-level |
| Apple M GPU | Architecture | Ada Lovelace |
| Custom | GPU Codename | AD107 (GN21-X2) |
| -Apple M3 Ultraor above | Recommended CPU | -Intel Core Ultra 7 155Hor above |
| 19491 | Steel Nomad Lite Score | 7949 |
| — | Time Spy | 8225 |
| 73690 | Solar Bay | 35461 |
| — | Port Royal | 4631 |
| — | Fire Strike | 21904 |
| 44759 | Wild Life Extreme | 15531 |
| — | Night Raid | 85428 |
| — | Base Clock | 1140-2070 MHz (configurable) |
| 1380 MHz | Boost Clock | 1605-2370 MHz (configurable) |
| 10240 | Shading Units | 2560 |
| 640 | Texture Mapping Units (TMUs) | 80 |
| 320 | Render Output Units (ROPs) | 32 |
| 1280 | Compute Units (Pipelines) | 20 |
| — | Tensor Cores | 80 |
| — | Ray-tracing Cores | 20 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 12MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |