Quick/Comparisons
50 | Gaming | 21 |
59 | Workstation | 17 |
41 | AI/ML |
Compare with something else?
Compare Apple M3 Ultra with another →12 |
61 | Energy Efficiency | 75 |
52 | Quick Comparison Final Score | 28 |
| Custom | Graphics Processor | GA107S |
| 10240 | Cores | 2048 |
| 640 / 320 | TMUs / ROPs | 64 / 40 |
| 7360.18 | Blender GPU | 489.72 |
| — | Forza Horizon 5 | — |
| — | The Witcher 3 | — |
| — | Counter-Strike 2 | — |
| — | Far Cry 6 | — |
| — | Hogwarts Legacy | — |
| — | Call of Duty | — |
| — | Ghost of Tsushima | — |
| — | Cyberpunk 2077 | — |
| — | Tomb Raider | — |
| — | Average FPS | — |
| — | 1080p High | — |
| — | 1080p Ultra | — |
| — | 1440p Ultra | — |
| — | 4K Ultra | — |
| — | Margin of Error | — |
| 147719 | GB6 Compute Score | 42999 |
| 204.5 img/sec | Background Blur | 77.2 img/sec |
| 126 img/sec | Face Detection | 40.3 img/sec |
| 6.53 Gpixels/sec | Horizon Detection | 1.65 Gpixels/sec |
| 10.7 Gpixels/sec | Edge Detection | 2.13 Gpixels/sec |
| 9.5 Gpixels/sec | Gaussian Blur | 2.04 Gpixels/sec |
| 1.05 Gpixels/sec | Feature Matching | 0.57 Gpixels/sec |
| 560.3 Gpixels/sec | Stereo Matching | 181.5 Gpixels/sec |
| 25267.8 FPS | Particle Physics | 5728.6 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 5795 |
| — | G2D Mark | 388 |
| — | DirectX 11 | 38 FPS |
| — | DirectX 12 | 30 FPS |
| — | GPU Compute | 2561 Ops/s |
| 21356 | GB6 ML Single Precision | 5997 |
| 23530 | GB6 ML Half Precision | 11591 |
| 21706 | GB6 ML Quantized | 4316 |
| 8094 | Image Classification (SP) | 2792 |
| 24595 | Image Segmentation (HP) | 5831 |
| 25190 | Image Super Resolution (Q) | 5115 |
| 38337 | Face Detection (HP) | 9491 |
| 111021 | Pose Estimation (Q) | 19218 |
| 2786 | Text Classification (SP) | 1996 |
| 4339 | Machine Translation (HP) | 1992 |
| 7668 | Object Detection (SP) | 2848 |
| 37488 | Depth Estimation (Q) | 10712 |
| 244483 | Style Transfer (SP) | 44344 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 442 GPixel/s | Pixel Fill Rate | 59 GPixel/s |
| 883 GTexel/s | Texture Fill Rate | 95 GTexel/s |
| 28.3 TFLOPS | FLOPS (FP32) | 6 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 4 GB |
| 6400 MHz | Memory Clock | 1500 MHz |
| — | Effective Memory Speed | 12000 Mbps |
| 1024-bit | Bus | 64-bit |
| No | ECC | — |
| 819.3 GB/s | Memory Bandwidth | 96 GB/s |
| Custom | Interface | PCIe 4.0 x8 |
| 140 W | TGP | 45 W |
| TSMC | Manufacturing | Samsung |
| 3 nm | Fabrication Process | 8 nm |
| — | Die Size | 200 mm² |
| 184 billion | Transistor Count | 8.7 billion |
| — | Transistor Density | 43.5 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | 8.6 |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| — | DisplayPort | 1.4 |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| March 5, 2025 | Released | January 1, 2022 |
| Desktop | Case | Laptop |
| Professional | Purpose | Gaming |
| High-end | Segment | Entry-level |
| Apple M GPU | Architecture | Ampere |
| Custom | GPU Codename | GA107S |
| -Apple M3 Ultraor above | Recommended CPU | -Intel Core i7 1255Uor above |
| 19491 | Steel Nomad Lite Score | — |
| — | Time Spy | — |
| 73690 | Solar Bay | — |
| — | Port Royal | — |
| — | Fire Strike | — |
| 44759 | Wild Life Extreme | — |
| — | Night Raid | — |
| — | Base Clock | 1155 MHz |
| 1380 MHz | Boost Clock | 1477 MHz |
| 10240 | Shading Units | 2048 |
| 640 | Texture Mapping Units (TMUs) | 64 |
| 320 | Render Output Units (ROPs) | 40 |
| 1280 | Compute Units (Pipelines) | 16 |
| — | Tensor Cores | 64 |
| — | Ray-tracing Cores | 16 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |