Quick/Comparisons
The GeForce RTX 5090 takes the lead in gaming performance with a perfect score of 100, while the Nvidia RTX Pro 6000 Blackwell closely follows with 99.
The Nvidia RTX Pro 6000 Blackwell outperforms the GeForce RTX 5090 in content creation with a score of 100, thanks to its superior memory capacity of 96GB.
The GeForce RTX 5090 takes the lead in AI/ML compute with a score of 95, narrowly edging out the Nvidia RTX Pro 6000 Blackwell with 93.
100 | Gaming | 99 |
93 | Workstation | 100 |
95 | AI/ML |
Compare with something else?
Compare GeForce RTX 5090 with another →93 |
49 | Energy Efficiency | 49 |
96 | Quick Comparison Final Score | 98 |
| GB202-300 | Graphics Processor | GB202-300-A1 |
| 21760 | Cores | 24064 |
| 680 / 176 | TMUs / ROPs | 752 / 190 |
| 15010.34 | Blender GPU | 16644.8 |
| 244 | Forza Horizon 5 | 230 |
| 323 | The Witcher 3 | 304 |
| 265 | Counter-Strike 2 | 259 |
| 202 | Far Cry 6 | 198 |
| 207 | Hogwarts Legacy | 192 |
| 327 | Call of Duty | 301 |
| 179 | Ghost of Tsushima | 172 |
| 229 | Cyberpunk 2077 | 221 |
| 299 | Tomb Raider | 303 |
| 253 | Average FPS | 242 |
| 294 | 1080p High | 282 |
| 253 | 1080p Ultra | 242 |
| 207 | 1440p Ultra | 197 |
| 124 | 4K Ultra | 118 |
| Low | Margin of Error | High |
| 419309 | GB6 Compute Score | 419470 |
| 333 img/sec | Background Blur | 366.1 img/sec |
| 298.5 img/sec | Face Detection | 248 img/sec |
| 21.7 Gpixels/sec | Horizon Detection | 23.3 Gpixels/sec |
| 33.7 Gpixels/sec | Edge Detection | 35.2 Gpixels/sec |
| 34.5 Gpixels/sec | Gaussian Blur | 45 Gpixels/sec |
| 2.53 Gpixels/sec | Feature Matching | 2.47 Gpixels/sec |
| 2850 Gpixels/sec | Stereo Matching | 3040 Gpixels/sec |
| 59146.9 FPS | Particle Physics | 42574.6 FPS |
| OpenCL | API | OpenCL |
| 38955 | G3D Mark Score | 37305 |
| 1413 | G2D Mark | 1375 |
| 341 FPS | DirectX 11 | 338 FPS |
| 178 FPS | DirectX 12 | 153 FPS |
| 24632 Ops/s | GPU Compute | 26895 Ops/s |
| 114685 | Face Detection (HP) | 103224 |
| 344729 | Pose Estimation (Q) | 338587 |
| 4464 | Text Classification (SP) | 3894 |
| 8184 | Machine Translation (HP) | 7291 |
| 27802 | Object Detection (SP) | 26498 |
| 85572 | Depth Estimation (Q) | 90853 |
| 687080 | Style Transfer (SP) | 635289 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 53861 | GB6 ML Single Precision | 50053 |
| 81080 | GB6 ML Half Precision | 74036 |
| 38938 | GB6 ML Quantized | 37636 |
| 18948 | Image Classification (SP) | 16903 |
| 55109 | Image Segmentation (HP) | 48518 |
| 58478 | Image Super Resolution (Q) | 54618 |
| 424 GPixel/s | Pixel Fill Rate | 497 GPixel/s |
| 1637 GTexel/s | Texture Fill Rate | 1968 GTexel/s |
| 104.8 TFLOPS | FLOPS (FP32) | 126 TFLOPS |
| GDDR7 | Memory Type | GDDR7 |
| 32 GB | Memory Size | 96 GB |
| 1750 MHz | Memory Clock | 1750 MHz |
| 28000 Mbps | Effective Memory Speed | 28000 Mbps |
| 512-bit | Bus | 512-bit |
| No | ECC | Yes |
| 1792 GB/s | Memory Bandwidth | 1792 GB/s |
| PCIe 5.0 x16 | Interface | PCIe 5.0 x16 |
| 575 W | TGP | 600 W |
| TSMC | Manufacturing | TSMC |
| 4 nm | Fabrication Process | 5 nm |
| 750 mm² | Die Size | 750 mm² |
| 92.2 billion | Transistor Count | 92.2 billion |
| 122.93 MTr/mm² | Transistor Density | 122.93 MTr/mm² |
| 12.2 | DirectX | 12.2 |
| 1.4 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 3 | OpenCL | 3 |
| 12 | CUDA | 12 |
| Yes | Ray Tracing | Yes |
| DLSS 4 | DLSS | DLSS 4 |
| 2.1b | DisplayPort | 2.1b |
| Nvidia | Vendor | Nvidia |
| Discrete | Build | Discrete |
| January 7, 2025 | Released | March 18, 2026 |
| Desktop | Case | Desktop |
| Gaming | Purpose | Professional |
| High-end | Segment | High-end |
| Blackwell 2.0 | Architecture | Blackwell 2.0 |
| GB202-300 | GPU Codename | GB202-300-A1 |
| -Intel Core Ultra 9 285Kor above | Recommended CPU | -AMD Ryzen 9 9955HX3Dor above |
| 52638 | Steel Nomad Lite Score | 50516 |
| 46941 | Time Spy | 42437 |
| 235224 | Solar Bay | 258808 |
| 38113 | Port Royal | 39952 |
| 87433 | Fire Strike | 72183 |
| 108736 | Wild Life Extreme | 110070 |
| 207058 | Night Raid | 216213 |
| 2017 MHz | Base Clock | 1590 MHz |
| 2407 MHz | Boost Clock | 2617 MHz |
| 21760 | Shading Units | 24064 |
| 680 | Texture Mapping Units (TMUs) | 752 |
| 176 | Render Output Units (ROPs) | 190 |
| 170 | Compute Units (Pipelines) | 188 |
| 680 | Tensor Cores | 752 |
| 170 | Ray-tracing Cores | 188 |
| 128KB per cluster | L1 Cache | 128KB per cluster |
| 96MB shared | L2 Cache | 128MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |