Quick/Comparisons
50 | Gaming | 100 |
59 | Workstation | 93 |
41 | AI/ML |
Compare with something else?
Compare Apple M3 Ultra with another →95 |
61 | Energy Efficiency | 49 |
52 | Quick Comparison Final Score | 96 |
| Custom | Graphics Processor | GB202-300 |
| 10240 | Cores | 21760 |
| 640 / 320 | TMUs / ROPs | 680 / 176 |
| 7360.18 | Blender GPU | 15010.34 |
| — | Forza Horizon 5 | 244 |
| — | The Witcher 3 | 323 |
| — | Counter-Strike 2 | 265 |
| — | Far Cry 6 | 202 |
| — | Hogwarts Legacy | 207 |
| — | Call of Duty | 327 |
| — | Ghost of Tsushima | 179 |
| — | Cyberpunk 2077 | 229 |
| — | Tomb Raider | 299 |
| — | Average FPS | 253 |
| — | 1080p High | 294 |
| — | 1080p Ultra | 253 |
| — | 1440p Ultra | 207 |
| — | 4K Ultra | 124 |
| — | Margin of Error | Low |
| 147719 | GB6 Compute Score | 419309 |
| 204.5 img/sec | Background Blur | 333 img/sec |
| 126 img/sec | Face Detection | 298.5 img/sec |
| 6.53 Gpixels/sec | Horizon Detection | 21.7 Gpixels/sec |
| 10.7 Gpixels/sec | Edge Detection | 33.7 Gpixels/sec |
| 9.5 Gpixels/sec | Gaussian Blur | 34.5 Gpixels/sec |
| 1.05 Gpixels/sec | Feature Matching | 2.53 Gpixels/sec |
| 560.3 Gpixels/sec | Stereo Matching | 2850 Gpixels/sec |
| 25267.8 FPS | Particle Physics | 59146.9 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 38955 |
| — | G2D Mark | 1413 |
| — | DirectX 11 | 341 FPS |
| — | DirectX 12 | 178 FPS |
| — | GPU Compute | 24632 Ops/s |
| 21356 | GB6 ML Single Precision | 53861 |
| 23530 | GB6 ML Half Precision | 81080 |
| 21706 | GB6 ML Quantized | 38938 |
| 8094 | Image Classification (SP) | 18948 |
| 24595 | Image Segmentation (HP) | 55109 |
| 25190 | Image Super Resolution (Q) | 58478 |
| 38337 | Face Detection (HP) | 114685 |
| 111021 | Pose Estimation (Q) | 344729 |
| 2786 | Text Classification (SP) | 4464 |
| 4339 | Machine Translation (HP) | 8184 |
| 7668 | Object Detection (SP) | 27802 |
| 37488 | Depth Estimation (Q) | 85572 |
| 244483 | Style Transfer (SP) | 687080 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 442 GPixel/s | Pixel Fill Rate | 424 GPixel/s |
| 883 GTexel/s | Texture Fill Rate | 1637 GTexel/s |
| 28.3 TFLOPS | FLOPS (FP32) | 104.8 TFLOPS |
| System Shared | Memory Type | GDDR7 |
| — | Memory Size | 32 GB |
| 6400 MHz | Memory Clock | 1750 MHz |
| — | Effective Memory Speed | 28000 Mbps |
| 1024-bit | Bus | 512-bit |
| No | ECC | No |
| 819.3 GB/s | Memory Bandwidth | 1792 GB/s |
| Custom | Interface | PCIe 5.0 x16 |
| 140 W | TGP | 575 W |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 4 nm |
| — | Die Size | 750 mm² |
| 184 billion | Transistor Count | 92.2 billion |
| — | Transistor Density | 122.93 MTr/mm² |
| — | DirectX | 12.2 |
| — | Vulkan | 1.4 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | 12 |
| Yes | Ray Tracing | Yes |
| No | DLSS | DLSS 4 |
| — | DisplayPort | 2.1b |
| Apple | Vendor | Nvidia |
| Integrated | Build | Discrete |
| March 5, 2025 | Released | January 7, 2025 |
| Desktop | Case | Desktop |
| Professional | Purpose | Gaming |
| High-end | Segment | High-end |
| Apple M GPU | Architecture | Blackwell 2.0 |
| Custom | GPU Codename | GB202-300 |
| -Apple M3 Ultraor above | Recommended CPU | -Intel Core Ultra 9 285Kor above |
| 19491 | Steel Nomad Lite Score | 52638 |
| — | Time Spy | 46941 |
| 73690 | Solar Bay | 235224 |
| — | Port Royal | 38113 |
| — | Fire Strike | 87433 |
| 44759 | Wild Life Extreme | 108736 |
| — | Night Raid | 207058 |
| — | Base Clock | 2017 MHz |
| 1380 MHz | Boost Clock | 2407 MHz |
| 10240 | Shading Units | 21760 |
| 640 | Texture Mapping Units (TMUs) | 680 |
| 320 | Render Output Units (ROPs) | 176 |
| 1280 | Compute Units (Pipelines) | 170 |
| — | Tensor Cores | 680 |
| — | Ray-tracing Cores | 170 |
| — | L1 Cache | 128KB per cluster |
| — | L2 Cache | 96MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |