Quick/Comparisons
The Radeon RX 7900 XTX leads by a significant margin of 37 points, thanks to its excellent gaming performance. Its 81 score far surpasses the Apple M5 Max GPU's 44.
Although the Apple M5 Max GPU scores 21 points higher in workstation tasks, the Radeon RX 7900 XTX's 75 score edges out the competition by 21 points. Its capable AI/ML compute also contributes to its win.
The Apple M5 Max GPU's 66 score in AI/ML tasks is just 8 points shy of the Radeon RX 7900 XTX's 74, resulting in a close but not decisive win for the Radeon. Its 8-point lead is too small to declare a clear winner.
44 | Gaming | 81 |
54 | Workstation | 75 |
66 | AI/ML | 74 |
100 | Energy Efficiency | 43 |
58 | Quick Comparison Final Score | 77 |
| Custom | Graphics Processor | Navi 31 |
| 5120 | Cores | 6144 |
| 320 / 160 | TMUs / ROPs | 384 / 192 |
| 7937.27 | Blender GPU | 4055.35 |
| — | Forza Horizon 5 | 174 |
| — | The Witcher 3 | 256 |
| — | Counter-Strike 2 | 232 |
| — | Far Cry 6 | 182 |
| — | Hogwarts Legacy | 137 |
| — | Call of Duty | 255 |
| — | Ghost of Tsushima | 138 |
| — | Cyberpunk 2077 | 181 |
| — | Tomb Raider | 277 |
| — | Average FPS | 204 |
| — | 1080p High | 241 |
| — | 1080p Ultra | 204 |
| — | 1440p Ultra | 159 |
| — | 4K Ultra | 94 |
| — | Margin of Error | Low |
| 146407 | GB6 Compute Score | 224686 |
| 257 img/sec | Background Blur | 423.9 img/sec |
| 182.1 img/sec | Face Detection | 270.8 img/sec |
| 5.4 Gpixels/sec | Horizon Detection | 9.23 Gpixels/sec |
| 8.33 Gpixels/sec | Edge Detection | 13.7 Gpixels/sec |
| 8.86 Gpixels/sec | Gaussian Blur | 11.1 Gpixels/sec |
| 1.28 Gpixels/sec | Feature Matching | 1.74 Gpixels/sec |
| 455.4 Gpixels/sec | Stereo Matching | 871 Gpixels/sec |
| 21658 FPS | Particle Physics | 29891.2 FPS |
| OpenCL | API | OpenCL |
| — | G3D Mark Score | 31406 |
| — | G2D Mark | 1285 |
| — | DirectX 11 | 356 FPS |
| — | DirectX 12 | 133 FPS |
| — | GPU Compute | 17768 Ops/s |
| 29452 | GB6 ML Single Precision | 42092 |
| 46343 | GB6 ML Half Precision | 50673 |
| 43305 | GB6 ML Quantized | 32874 |
| 10711 | Image Classification (SP) | 18608 |
| 42406 | Image Segmentation (HP) | 25385 |
| 42905 | Image Super Resolution (Q) | 45641 |
| 73489 | Face Detection (HP) | 71321 |
| 289766 | Pose Estimation (Q) | 257258 |
| 6087 | Text Classification (SP) | 5240 |
| 8881 | Machine Translation (HP) | 7425 |
| 10130 | Object Detection (SP) | 21335 |
| 84224 | Depth Estimation (Q) | 66140 |
| 264552 | Style Transfer (SP) | 483867 |
| Core ML | Framework | ONNX |
| GPU | Backend | DirectML |
| 259 GPixel/s | Pixel Fill Rate | 480 GPixel/s |
| 518 GTexel/s | Texture Fill Rate | 959 GTexel/s |
| 16.6 TFLOPS | FLOPS (FP32) | 61.4 TFLOPS |
| System Shared | Memory Type | GDDR6 |
| — | Memory Size | 24 GB |
| 9600 MHz | Memory Clock | 2500 MHz |
| — | Effective Memory Speed | 20000 Mbps |
| 512-bit | Bus | 384-bit |
| No | ECC | No |
| 614 GB/s | Memory Bandwidth | 960 GB/s |
| Custom | Interface | PCIe 4.0 x16 |
| — | TGP | 355 W |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 5 nm |
| — | Die Size | 529 mm² |
| 28 billion | Transistor Count | 57 billion |
| — | Transistor Density | 107.75 MTr/mm² |
| — | DirectX | 12 |
| — | Vulkan | 1.3 |
| — | OpenGL | 4.6 |
| — | OpenCL | 2.2 |
| — | CUDA | — |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| — | DisplayPort | 2.1 |
| Apple | Vendor | Amd |
| Integrated | Build | Discrete |
| March 3, 2026 | Released | November 3, 2022 |
| Laptop | Case | Desktop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Apple M GPU | Architecture | RDNA 3.0 |
| Custom | GPU Codename | Navi 31 |
| -Apple M5 Max (40-сore GPU)or above | Recommended CPU | -Intel Core i7 13700Kor above |
| 17913 | Steel Nomad Lite Score | 29870 |
| — | Time Spy | 30766 |
| 79412 | Solar Bay | 123758 |
| — | Port Royal | 16749 |
| — | Fire Strike | 73217 |
| 42244 | Wild Life Extreme | 58530 |
| — | Night Raid | 184767 |
| — | Base Clock | 1929 MHz |
| 1620 MHz | Boost Clock | 2498 MHz |
| 5120 | Shading Units | 6144 |
| 320 | Texture Mapping Units (TMUs) | 384 |
| 160 | Render Output Units (ROPs) | 192 |
| 640 | Compute Units (Pipelines) | — |
| — | Tensor Cores | — |
| — | Ray-tracing Cores | 96 |
| — | L1 Cache | 256KB per cluster |
| — | L2 Cache | 6MB shared |
| 2 IPC | Instructions Per Cycle | 4 IPC |
Compare with something else?
Compare Apple M5 Max with another →