Quick/Comparisons
The Nvidia RTX 6000 Ada takes the lead in gaming performance, with a 4-point advantage over the GeForce RTX 5070 Ti.
The Nvidia RTX 6000 Ada dominates in content creation, with a 20-point gap in workstation performance and 48GB of GDDR6 memory.
The Nvidia RTX 6000 Ada edges out the GeForce RTX 5070 Ti in AI/ML compute, with a 9-point advantage and capable performance.
76 | Gaming | 80 |
74 | Workstation | 94 |
66 | AI/ML |
Compare with something else?
Compare GeForce RTX 5070 with another →75 |
49 | Energy Efficiency | 44 |
74 | Quick Comparison Final Score | 83 |
| GB203-200-A1 | Graphics Processor | AD102 |
| 8960 | Cores | 18176 |
| 280 / 96 | TMUs / ROPs | 568 / 192 |
| 7584.67 | Blender GPU | 11038.91 |
| 166 | Forza Horizon 5 | 165 |
| 213 | The Witcher 3 | 219 |
| 219 | Counter-Strike 2 | 221 |
| 159 | Far Cry 6 | 164 |
| 135 | Hogwarts Legacy | 133 |
| 205 | Call of Duty | 216 |
| 121 | Ghost of Tsushima | 123 |
| 174 | Cyberpunk 2077 | 172 |
| 260 | Tomb Raider | 267 |
| 184 | Average FPS | 187 |
| 217 | 1080p High | 219 |
| 184 | 1080p Ultra | 187 |
| 148 | 1440p Ultra | 148 |
| 86 | 4K Ultra | 89 |
| Low | Margin of Error | High |
| 240508 | GB6 Compute Score | 389848 |
| 297.9 img/sec | Background Blur | 618.3 img/sec |
| 218.6 img/sec | Face Detection | 399.9 img/sec |
| 10.1 Gpixels/sec | Horizon Detection | 15.1 Gpixels/sec |
| 13.9 Gpixels/sec | Edge Detection | 22.1 Gpixels/sec |
| 14.8 Gpixels/sec | Gaussian Blur | 26.3 Gpixels/sec |
| 2.03 Gpixels/sec | Feature Matching | 2.33 Gpixels/sec |
| 1290 Gpixels/sec | Stereo Matching | 2210 Gpixels/sec |
| 35438.3 FPS | Particle Physics | 53494.5 FPS |
| OpenCL | API | OpenCL |
| 32425 | G3D Mark Score | 28613 |
| 1329 | G2D Mark | 1005 |
| 304 FPS | DirectX 11 | 284 FPS |
| 126 FPS | DirectX 12 | 102 FPS |
| 18089 Ops/s | GPU Compute | 22439 Ops/s |
| 36504 | GB6 ML Single Precision | 37270 |
| 64459 | GB6 ML Half Precision | 53469 |
| 27312 | GB6 ML Quantized | 27850 |
| 14913 | Image Classification (SP) | 13367 |
| 45199 | Image Segmentation (HP) | 32477 |
| 39360 | Image Super Resolution (Q) | 42056 |
| 84188 | Face Detection (HP) | 69004 |
| 161400 | Pose Estimation (Q) | 230965 |
| 4010 | Text Classification (SP) | 3491 |
| 7973 | Machine Translation (HP) | 5545 |
| 18934 | Object Detection (SP) | 18668 |
| 67761 | Depth Estimation (Q) | 68370 |
| 335879 | Style Transfer (SP) | 520389 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 235 GPixel/s | Pixel Fill Rate | 481 GPixel/s |
| 687 GTexel/s | Texture Fill Rate | 1423 GTexel/s |
| 43.9 TFLOPS | FLOPS (FP32) | 91.1 TFLOPS |
| GDDR7 | Memory Type | GDDR6 |
| 16 GB | Memory Size | 48 GB |
| 1750 MHz | Memory Clock | 2500 MHz |
| 28000 Mbps | Effective Memory Speed | 20000 Mbps |
| 256-bit | Bus | 384-bit |
| No | ECC | No |
| 896 GB/s | Memory Bandwidth | 960 GB/s |
| PCIe 5.0 x16 | Interface | PCIe 4.0 x16 |
| 300 W | TGP | 300 W |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 5 nm |
| 378 mm² | Die Size | 609 mm² |
| 45.6 billion | Transistor Count | 76 billion |
| 120.63 MTr/mm² | Transistor Density | 124.79 MTr/mm² |
| 12.2 | DirectX | 12 |
| 1.4 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 12 | CUDA | 8.9 |
| Yes | Ray Tracing | Yes |
| DLSS 4 | DLSS | DLSS 3 |
| 2.1b | DisplayPort | 1.4a |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| January 7, 2025 | Released | December 3, 2022 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Professional |
| Mid-range | Segment | High-end |
| Blackwell 2.0 | Architecture | Ada Lovelace |
| GB203-200-A1 | GPU Codename | AD102 |
| -Intel Core Ultra 5 245Kor above | Recommended CPU | -Intel Core i9 14900Kor above |
| 29989 | Steel Nomad Lite Score | 32284 |
| 27815 | Time Spy | 26686 |
| 136964 | Solar Bay | 145429 |
| 19613 | Port Royal | 19557 |
| 67382 | Fire Strike | 62803 |
| 57734 | Wild Life Extreme | 53640 |
| 189734 | Night Raid | 148225 |
| 2295 MHz | Base Clock | 915 MHz |
| 2452 MHz | Boost Clock | 2505 MHz |
| 8960 | Shading Units | 18176 |
| 280 | Texture Mapping Units (TMUs) | 568 |
| 96 | Render Output Units (ROPs) | 192 |
| 70 | Compute Units (Pipelines) | 142 |
| 280 | Tensor Cores | 568 |
| 70 | Ray-tracing Cores | 142 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 48MB shared | L2 Cache | 96MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |