Quick/Comparisons
The GeForce RTX 5070 Ti's 76 score in gaming significantly outpaces the Radeon RX 6650 XT's 30, thanks to its Blackwell 2.0 architecture and 16GB GDDR7 memory.
The GeForce RTX 5070 Ti takes the lead in content creation with a 74 score, driven by its solid workstation performance and large 16GB memory pool.
The GeForce RTX 5070 Ti's 66 score in AI/ML is higher than the Radeon RX 6650 XT's 31, thanks to its dedicated hardware ray tracing support and AI tensor cores.
76 | Gaming | 30 |
74 | Workstation | 34 |
66 | AI/ML | 31 |
49 | Energy Efficiency | 39 |
74 | Quick Comparison Final Score | 31 |
| GB203-200-A1 | Graphics Processor | Navi 23 |
| 8960 | Cores | 2048 |
| 280 / 96 | TMUs / ROPs | 128 / 64 |
| 7584.67 | Blender GPU | 1138.24 |
| 166 | Forza Horizon 5 | 76 |
| 213 | The Witcher 3 | 101 |
| 219 | Counter-Strike 2 | 130 |
| 159 | Far Cry 6 | 93 |
| 135 | Hogwarts Legacy | 59 |
| 205 | Call of Duty | 87 |
| 121 | Ghost of Tsushima | 63 |
| 174 | Cyberpunk 2077 | 65 |
| 260 | Tomb Raider | 111 |
| 184 | Average FPS | 87 |
| 217 | 1080p High | 112 |
| 184 | 1080p Ultra | 87 |
| 148 | 1440p Ultra | 61 |
| 86 | 4K Ultra | 32 |
| Low | Margin of Error | Low |
| 240508 | GB6 Compute Score | 90831 |
| 297.9 img/sec | Background Blur | 220.1 img/sec |
| 218.6 img/sec | Face Detection | 128.3 img/sec |
| 10.1 Gpixels/sec | Horizon Detection | 2.94 Gpixels/sec |
| 13.9 Gpixels/sec | Edge Detection | 4.13 Gpixels/sec |
| 14.8 Gpixels/sec | Gaussian Blur | 3.68 Gpixels/sec |
| 2.03 Gpixels/sec | Feature Matching | 0.94 Gpixels/sec |
| 1290 Gpixels/sec | Stereo Matching | 408.2 Gpixels/sec |
| 35438.3 FPS | Particle Physics | 10694.7 FPS |
| OpenCL | API | OpenCL |
| 32425 | G3D Mark Score | 17117 |
| 1329 | G2D Mark | 924 |
| 304 FPS | DirectX 11 | 168 FPS |
| 126 FPS | DirectX 12 | 62 FPS |
| 18089 Ops/s | GPU Compute | 7461 Ops/s |
| 36504 | GB6 ML Single Precision | 17945 |
| 64459 | GB6 ML Half Precision | 21242 |
| 27312 | GB6 ML Quantized | 13940 |
| 14913 | Image Classification (SP) | 8220 |
| 45199 | Image Segmentation (HP) | 9263 |
| 39360 | Image Super Resolution (Q) | 20723 |
| 84188 | Face Detection (HP) | 27074 |
| 161400 | Pose Estimation (Q) | 76541 |
| 4010 | Text Classification (SP) | 2796 |
| 7973 | Machine Translation (HP) | 4281 |
| 18934 | Object Detection (SP) | 6880 |
| 67761 | Depth Estimation (Q) | 33354 |
| 335879 | Style Transfer (SP) | 185358 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 235 GPixel/s | Pixel Fill Rate | 169 GPixel/s |
| 687 GTexel/s | Texture Fill Rate | 337 GTexel/s |
| 43.9 TFLOPS | FLOPS (FP32) | 10.8 TFLOPS |
| GDDR7 | Memory Type | GDDR6 |
| 16 GB | Memory Size | 8 GB |
| 1750 MHz | Memory Clock | 2190 MHz |
| 28000 Mbps | Effective Memory Speed | 17500 Mbps |
| 256-bit | Bus | 128-bit |
| No | ECC | No |
| 896 GB/s | Memory Bandwidth | 280.3 GB/s |
| PCIe 5.0 x16 | Interface | PCIe 4.0 x8 |
| 300 W | TGP | 176 W |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 7 nm |
| 378 mm² | Die Size | 237 mm² |
| 45.6 billion | Transistor Count | 11 billion |
| 120.63 MTr/mm² | Transistor Density | 46.41 MTr/mm² |
| 12.2 | DirectX | 12 |
| 1.4 | Vulkan | 1.3 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 2.1 |
| 12 | CUDA | — |
| Yes | Ray Tracing | Yes |
| DLSS 4 | DLSS | No |
| 2.1b | DisplayPort | 1.4a |
| Nvidia | Vendor | Amd |
| Discrete | Build | Discrete |
| January 7, 2025 | Released | May 10, 2022 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| Mid-range | Segment | High-end |
| Blackwell 2.0 | Architecture | RDNA 2.0 |
| GB203-200-A1 | GPU Codename | Navi 23 |
| -Intel Core Ultra 5 245Kor above | Recommended CPU | -Intel Core i5 12600Kor above |
| 29989 | Steel Nomad Lite Score | 9265 |
| 27815 | Time Spy | 9897 |
| 136964 | Solar Bay | 38892 |
| 19613 | Port Royal | 4897 |
| 67382 | Fire Strike | 29357 |
| 57734 | Wild Life Extreme | 18034 |
| 189734 | Night Raid | 116256 |
| 2295 MHz | Base Clock | 2055 MHz |
| 2452 MHz | Boost Clock | 2635 MHz |
| 8960 | Shading Units | 2048 |
| 280 | Texture Mapping Units (TMUs) | 128 |
| 96 | Render Output Units (ROPs) | 64 |
| 70 | Compute Units (Pipelines) | — |
| 280 | Tensor Cores | No |
| 70 | Ray-tracing Cores | 32 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 48MB shared | L2 Cache | 2MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |
Compare with something else?
Compare GeForce RTX 5070 with another →