Quick/Comparisons
50 | Gaming | 62 |
59 | Workstation | 68 |
41 | AI/ML |
Compare with something else?
Compare Apple M3 Ultra with another →52 |
61 | Energy Efficiency | 45 |
52 | Quick Comparison Final Score | 62 |
| Custom | Graphics Processor | AD103 |
| 10240 | Cores | 7680 |
| 640 / 320 | TMUs / ROPs | 240 / 80 |
| 7360.18 | Blender GPU | 5579.97 |
| — | Forza Horizon 5 | 126 |
| — | The Witcher 3 | 175 |
| — | Counter-Strike 2 | 213 |
| — | Far Cry 6 | 145 |
| — | Hogwarts Legacy | 106 |
| — | Call of Duty | 160 |
| — | Ghost of Tsushima | 99 |
| — | Cyberpunk 2077 | 134 |
| — | Tomb Raider | 216 |
| — | Average FPS | 153 |
| — | 1080p High | 181 |
| — | 1080p Ultra | 153 |
| — | 1440p Ultra | 119 |
| — | 4K Ultra | 68 |
| — | Margin of Error | High |
| 147719 | GB6 Compute Score | 206673 |
| 204.5 img/sec | Background Blur | 309.9 img/sec |
| 126 img/sec | Face Detection | 198.4 img/sec |
| 6.53 Gpixels/sec | Horizon Detection | 7.42 Gpixels/sec |
| 10.7 Gpixels/sec | Edge Detection | 9.94 Gpixels/sec |
| 9.5 Gpixels/sec | Gaussian Blur | 12.4 Gpixels/sec |
| 1.05 Gpixels/sec | Feature Matching | 2.02 Gpixels/sec |
| 560.3 Gpixels/sec | Stereo Matching | 1060 Gpixels/sec |
| 25267.8 FPS | Particle Physics | 31078.1 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 28210 |
| — | G2D Mark | 1183 |
| — | DirectX 11 | 258 FPS |
| — | DirectX 12 | 92 FPS |
| — | GPU Compute | 16666 Ops/s |
| 21356 | GB6 ML Single Precision | 26396 |
| 23530 | GB6 ML Half Precision | 40553 |
| 21706 | GB6 ML Quantized | 20133 |
| 8094 | Image Classification (SP) | 11193 |
| 24595 | Image Segmentation (HP) | 20124 |
| 25190 | Image Super Resolution (Q) | 34904 |
| 38337 | Face Detection (HP) | 42820 |
| 111021 | Pose Estimation (Q) | 149991 |
| 2786 | Text Classification (SP) | 3789 |
| 4339 | Machine Translation (HP) | 4268 |
| 7668 | Object Detection (SP) | 13068 |
| 37488 | Depth Estimation (Q) | 47003 |
| 244483 | Style Transfer (SP) | 300643 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 442 GPixel/s | Pixel Fill Rate | 206 GPixel/s |
| 883 GTexel/s | Texture Fill Rate | 619 GTexel/s |
| 28.3 TFLOPS | FLOPS (FP32) | 39.6 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 24 GB |
| 6400 MHz | Memory Clock | 2250 MHz |
| — | Effective Memory Speed | 18000 Mbps |
| 1024-bit | Bus | 192-bit |
| No | ECC | No |
| 819.3 GB/s | Memory Bandwidth | 432 GB/s |
| Custom | Interface | PCIe 4.0 x16 |
| 140 W | TGP | 210 W |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 5 nm |
| — | Die Size | 379 mm² |
| 184 billion | Transistor Count | 45 billion |
| — | Transistor Density | 118.73 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 | August 23, 2023 |
| Desktop | Case | Desktop |
| Professional | Purpose | Professional |
| High-end | Segment | High-end |
| Apple M GPU | Architecture | Ada Lovelace |
| Custom | GPU Codename | AD103 |
| -Apple M3 Ultraor above | Recommended CPU | -Intel Core i7 14700Kor above |
| 19491 | Steel Nomad Lite Score | 21566 |
| — | Time Spy | 19844 |
| 73690 | Solar Bay | 95907 |
| — | Port Royal | 12727 |
| — | Fire Strike | 48020 |
| 44759 | Wild Life Extreme | 40308 |
| — | Night Raid | — |
| — | Base Clock | 2070 MHz |
| 1380 MHz | Boost Clock | 2580 MHz |
| 10240 | Shading Units | 7680 |
| 640 | Texture Mapping Units (TMUs) | 240 |
| 320 | Render Output Units (ROPs) | 80 |
| 1280 | Compute Units (Pipelines) | 60 |
| — | Tensor Cores | 240 |
| — | Ray-tracing Cores | 60 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 48MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |