Quick/Comparisons
The GeForce RTX 5090's exceptional gaming performance is unmatched, thanks to its 100 score in this category. Its Blackwell 2.0 architecture and 32GB GDDR7 memory provide a significant edge over the Radeon RX 6800.
The GeForce RTX 5090's strong workstation performance and 93 score give it a clear advantage in content creation tasks, making it the better choice for professionals. Its Blackwell 2.0 architecture and PCIe 5.0 x16 interface provide a solid foundation for demanding workloads.
The GeForce RTX 5090's capable AI/ML compute and 95 score make it the better choice for AI and machine learning tasks, thanks to its strong performance in this category. Its Blackwell 2.0 architecture and 32GB GDDR7 memory provide a significant edge over the Radeon RX 6800.
100 | Gaming | 45 |
93 | Workstation | 47 |
95 | AI/ML |
Compare with something else?
Compare GeForce RTX 5090 with another →44 |
49 | Energy Efficiency | 36 |
96 | Quick Comparison Final Score | 45 |
| GB202-300 | Graphics Processor | Navi 21 |
| 21760 | Cores | 3840 |
| 680 / 176 | TMUs / ROPs | 240 / 96 |
| 15010.34 | Blender GPU | 1894.91 |
| 244 | Forza Horizon 5 | 116 |
| 323 | The Witcher 3 | 152 |
| 265 | Counter-Strike 2 | 186 |
| 202 | Far Cry 6 | 125 |
| 207 | Hogwarts Legacy | 88 |
| 327 | Call of Duty | 134 |
| 179 | Ghost of Tsushima | 90 |
| 229 | Cyberpunk 2077 | 98 |
| 299 | Tomb Raider | 168 |
| 253 | Average FPS | 129 |
| 294 | 1080p High | 157 |
| 253 | 1080p Ultra | 129 |
| 207 | 1440p Ultra | 95 |
| 124 | 4K Ultra | 55 |
| Low | Margin of Error | Low |
| 419309 | GB6 Compute Score | 135060 |
| 333 img/sec | Background Blur | 298.6 img/sec |
| 298.5 img/sec | Face Detection | 186.2 img/sec |
| 21.7 Gpixels/sec | Horizon Detection | 4.75 Gpixels/sec |
| 33.7 Gpixels/sec | Edge Detection | 6.68 Gpixels/sec |
| 34.5 Gpixels/sec | Gaussian Blur | 5.58 Gpixels/sec |
| 2.53 Gpixels/sec | Feature Matching | 1.16 Gpixels/sec |
| 2850 Gpixels/sec | Stereo Matching | 579 Gpixels/sec |
| 59146.9 FPS | Particle Physics | 18755.1 FPS |
| OpenCL | API | OpenCL |
| 38955 | G3D Mark Score | 22023 |
| 1413 | G2D Mark | 998 |
| 341 FPS | DirectX 11 | 212 FPS |
| 178 FPS | DirectX 12 | 89 FPS |
| 24632 Ops/s | GPU Compute | 10764 Ops/s |
| 114685 | Face Detection (HP) | 38124 |
| 344729 | Pose Estimation (Q) | 119493 |
| 4464 | Text Classification (SP) | 2425 |
| 8184 | Machine Translation (HP) | 5228 |
| 27802 | Object Detection (SP) | 11986 |
| 85572 | Depth Estimation (Q) | 46083 |
| 687080 | Style Transfer (SP) | 281860 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 53861 | GB6 ML Single Precision | 25097 |
| 81080 | GB6 ML Half Precision | 27988 |
| 38938 | GB6 ML Quantized | 19116 |
| 18948 | Image Classification (SP) | 11245 |
| 55109 | Image Segmentation (HP) | 13617 |
| 58478 | Image Super Resolution (Q) | 27747 |
| 424 GPixel/s | Pixel Fill Rate | 202 GPixel/s |
| 1637 GTexel/s | Texture Fill Rate | 505 GTexel/s |
| 104.8 TFLOPS | FLOPS (FP32) | 16.2 TFLOPS |
| GDDR7 | Memory Type | GDDR6 |
| 32 GB | Memory Size | 16 GB |
| 1750 MHz | Memory Clock | 2000 MHz |
| 28000 Mbps | Effective Memory Speed | 16000 Mbps |
| 512-bit | Bus | 256-bit |
| No | ECC | No |
| 1792 GB/s | Memory Bandwidth | 512 GB/s |
| PCIe 5.0 x16 | Interface | PCIe 4.0 x16 |
| 575 W | TGP | 250 W |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 7 nm |
| 750 mm² | Die Size | 520 mm² |
| 92.2 billion | Transistor Count | 26 billion |
| 122.93 MTr/mm² | Transistor Density | 50 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 | October 28, 2020 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Gaming |
| High-end | Segment | High-end |
| Blackwell 2.0 | Architecture | RDNA 2.0 |
| GB202-300 | GPU Codename | Navi 21 |
| -Intel Core Ultra 9 285Kor above | Recommended CPU | -Intel Core i5 12600or above |
| 52638 | Steel Nomad Lite Score | 14548 |
| 46941 | Time Spy | 16220 |
| 235224 | Solar Bay | 64755 |
| 38113 | Port Royal | 8137 |
| 87433 | Fire Strike | 44377 |
| 108736 | Wild Life Extreme | 29764 |
| 207058 | Night Raid | 142721 |
| 2017 MHz | Base Clock | 1700 MHz |
| 2407 MHz | Boost Clock | 2105 MHz |
| 21760 | Shading Units | 3840 |
| 680 | Texture Mapping Units (TMUs) | 240 |
| 176 | Render Output Units (ROPs) | 96 |
| 170 | Compute Units (Pipelines) | — |
| 680 | Tensor Cores | No |
| 170 | Ray-tracing Cores | 60 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 96MB shared | L2 Cache | 4MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |