Quick/Comparisons
A mid-to-upper-range Nvidia GeForce RTX 4070 Ti SUPER graphics card with capable workstation performance (72/100). Suited for 1440p/1080p gaming and content creation.
| Gaming | 70 /100 |
| Workstation | 72 /100 |
| AI/ML | 59 /100 |
| Energy Efficiency | 46 |
Looking for more comparisons?
Compare GeForce RTX →| Quick Comparison Final Score | 69 /100 |
| Graphics Processor | AD103 |
| Cores | 8448 |
| TMUs / ROPs | 264 / 96 |
| Blender GPU | 7089.35 |
| Forza Horizon 5 | 150 |
| The Witcher 3 | 198 |
| Counter-Strike 2 | 220 |
| Far Cry 6 | 152 |
| Hogwarts Legacy | 118 |
| Call of Duty | 188 |
| Ghost of Tsushima | 111 |
| Cyberpunk 2077 | 155 |
| Tomb Raider | 249 |
| Average FPS | 171 |
| 1080p High | 200 |
| 1080p Ultra | 171 |
| 1440p Ultra | 134 |
| 4K Ultra | 79 |
| Margin of Error | Low |
| GB6 Compute Score | 237068 |
| Background Blur | 331.7 img/sec |
| Face Detection | 211.6 img/sec |
| Horizon Detection | 9.63 Gpixels/sec |
| Edge Detection | 13.7 Gpixels/sec |
| Gaussian Blur | 13.9 Gpixels/sec |
| Feature Matching | 2.15 Gpixels/sec |
| Stereo Matching | 1210 Gpixels/sec |
| Particle Physics | 33566 FPS |
| API | OpenCL |
| G3D Mark Score | 31798 |
| G2D Mark | 1240 |
| DirectX 11 | 277 FPS |
| DirectX 12 | 119 FPS |
| GPU Compute | 18264 Ops/s |
| GB6 ML Single Precision | 32793 |
| GB6 ML Half Precision | 47412 |
| GB6 ML Quantized | 24566 |
| Image Classification (SP) | 12970 |
| Image Segmentation (HP) | 26364 |
| Image Super Resolution (Q) | 37217 |
| Face Detection (HP) | 58720 |
| Pose Estimation (Q) | 177604 |
| Text Classification (SP) | 3815 |
| Machine Translation (HP) | 5829 |
| Object Detection (SP) | 15675 |
| Depth Estimation (Q) | 62447 |
| Style Transfer (SP) | 357918 |
| Framework | ONNX |
| Backend | DirectML |
| Pixel Fill Rate | 251 GPixel/s |
| Texture Fill Rate | 689 GTexel/s |
| FLOPS (FP32) | 44.1 TFLOPS |
| Memory Type | GDDR6X |
| Memory Size | 16 GB |
| Memory Clock | 1313 MHz |
| Effective Memory Speed | 21000 Mbps |
| Bus | 256-bit |
| ECC | No |
| Memory Bandwidth | 672.3 GB/s |
| Interface | PCIe 4.0 x16 |
| TGP | 285 W |
| Manufacturing | TSMC |
| Fabrication Process | 5 nm |
| Die Size | 379 mm² |
| Transistor Count | 45 billion |
| Transistor Density | 118.73 MTr/mm² |
| DirectX | 12 |
| Vulkan | 1.3 |
| OpenGL | 4.6 |
| OpenCL | 3 |
| CUDA | 8.9 |
| Ray Tracing | Yes |
| DLSS | DLSS 3 |
| DisplayPort | 1.4a |
| Vendor | Nvidia |
| Build | Discrete |
| Released | January 24, 2024 |
| Case | Desktop |
| Purpose | Gaming |
| Segment | High-end |
| Architecture | Ada Lovelace |
| GPU Codename | AD103 |
| Recommended CPU | -Intel Core i7 14700Kor above |
| Steel Nomad Lite Score | 25741 |
| Time Spy | 24222 |
| Solar Bay | 116542 |
| Port Royal | 15802 |
| Fire Strike | 56340 |
| Wild Life Extreme | 49901 |
| Night Raid | 176609 |
| Base Clock | 2340 MHz |
| Boost Clock | 2610 MHz |
| Shading Units | 8448 |
| Texture Mapping Units (TMUs) | 264 |
| Render Output Units (ROPs) | 96 |
| Compute Units (Pipelines) | 66 |
| Tensor Cores | 264 |
| Ray-tracing Cores | 66 |
| L1 Cache | 128KB per cluster |
| L2 Cache | 48MB shared |
| Instructions Per Cycle | 2 IPC |
84/100