Quick/Comparisons
14 | Gaming | 16 |
24 | Workstation | 25 |
| — | AI/ML |
Compare with something else?
Compare Apple M4 GPU with another →97 | Energy Efficiency | 100 |
32 | Quick Comparison Final Score | 33 |
| Custom | Graphics Processor | Panther Lake |
| 1024 | Cores | 1536 |
| 64 / 32 | TMUs / ROPs | 48 / 24 |
| 1060.26 | Blender GPU | 1268.79 |
| — | Forza Horizon 5 | 38 |
| — | The Witcher 3 | 55 |
| — | Counter-Strike 2 | 60 |
| — | Far Cry 6 | 49 |
| — | Hogwarts Legacy | 21 |
| — | Call of Duty | 32 |
| — | Ghost of Tsushima | 25 |
| — | Cyberpunk 2077 | 33 |
| — | Tomb Raider | 52 |
| — | Average FPS | 41 |
| — | 1080p High | 52 |
| — | 1080p Ultra | 41 |
| — | 1440p Ultra | 29 |
| — | 4K Ultra | 14 |
| — | Margin of Error | Low |
| 30791 | GB6 Compute Score | 64546 |
| 57.4 img/sec | Background Blur | — |
| 41.3 img/sec | Face Detection | — |
| 1.21 Gpixels/sec | Horizon Detection | — |
| 1.74 Gpixels/sec | Edge Detection | — |
| 1.12 Gpixels/sec | Gaussian Blur | — |
| 0.43 Gpixels/sec | Feature Matching | — |
| 94.9 Gpixels/sec | Stereo Matching | — |
| 3902.5 FPS | Particle Physics | — |
| OpenCL | API | — |
| — | G3D Mark Score | 9029 |
| — | G2D Mark | 786 |
| — | DirectX 11 | 70 FPS |
| — | DirectX 12 | 50 FPS |
| — | GPU Compute | 3702 Ops/s |
| — | GB6 ML Single Precision | 11871 |
| — | GB6 ML Half Precision | 20120 |
| — | GB6 ML Quantized | 7879 |
| — | Image Classification (SP) | 5487 |
| — | Image Segmentation (HP) | 16864 |
| — | Image Super Resolution (Q) | 8788 |
| — | Face Detection (HP) | 25029 |
| — | Pose Estimation (Q) | 18852 |
| — | Text Classification (SP) | 2115 |
| — | Machine Translation (HP) | 4266 |
| — | Object Detection (SP) | 7449 |
| — | Depth Estimation (Q) | 22207 |
| — | Style Transfer (SP) | 95039 |
| — | Framework | ONNX |
| — | Backend | DirectML |
| 47 GPixel/s | Pixel Fill Rate | 60 GPixel/s |
| 94 GTexel/s | Texture Fill Rate | 120 GTexel/s |
| 3 TFLOPS | FLOPS (FP32) | 7.7 TFLOPS |
| System Shared | Memory Type | System Shared |
| — | Memory Size | — |
| 7500 MHz | Memory Clock | — |
| 15000 Mbps | Effective Memory Speed | — |
| 128-bit | Bus | — |
| No | ECC | No |
| 120 GB/s | Memory Bandwidth | — |
| Custom | Interface | PCIe 5.0 |
| 15 W | TGP | — |
| TSMC | Manufacturing | TSMC |
| 3 nm | Fabrication Process | 3 nm |
| — | Die Size | — |
| — | Transistor Count | — |
| — | Transistor Density | — |
| — | DirectX | 12.1 |
| — | Vulkan | 1.4 |
| — | OpenGL | 4.6 |
| — | OpenCL | 3 |
| — | CUDA | — |
| Yes | Ray Tracing | Yes |
| No | DLSS | No |
| — | DisplayPort | 2.1 |
| Apple | Vendor | Intel |
| Integrated | Build | Integrated |
| October 29, 2024 | Released | January 5, 2026 |
| Laptop | Case | Laptop |
| Professional | Purpose | Gaming |
| Mid-range | Segment | Mid-range |
| Apple M GPU | Architecture | Celestial |
| Custom | GPU Codename | Panther Lake |
| -Apple M4 (8-Core)or above | Recommended CPU | -Intel Core Ultra X9 388Hor above |
| 3450 | Steel Nomad Lite Score | 4242 |
| — | Time Spy | 6595 |
| 13908 | Solar Bay | 26399 |
| — | Port Royal | 3886 |
| — | Fire Strike | 16770 |
| 7928 | Wild Life Extreme | 11707 |
| — | Night Raid | 68498 |
| 500 MHz | Base Clock | 300 MHz |
| 1470 MHz | Boost Clock | 2500 MHz |
| 1024 | Shading Units | 1536 |
| 64 | Texture Mapping Units (TMUs) | 48 |
| 32 | Render Output Units (ROPs) | 24 |
| 128 | Compute Units (Pipelines) | 12 |
| — | Tensor Cores | — |
| — | Ray-tracing Cores | — |
| — | L1 Cache | 64KB per cluster |
| — | L2 Cache | 16MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |