GeForce RTX 5090 vs Nvidia RTX Pro 6000 Blackwell
[ VERDICT ]
Winner: Nvidia RTX Pro 6000 Blackwell (98/100)
Nvidia RTX Pro 6000 Blackwell leads overall with a score of 98/100 compared to GeForce RTX 5090 (96/100), offering superior measured performance and technical specifications.
Benchmark Performance Comparison
Normalized Lab Scores| Benchmark | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Efficiency | 48.0 Poor | 49.0 Poor WINNER |
| Gaming | 100.0 Excellent WINNER | 99.0 Excellent |
Detailed Technical Specifications
3D Mark
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Fire Strike | 87569 | 75325 |
| Night Raid | 206648 | 214655 |
| Port Royal | 38202 | 39888 |
| Solar Bay | 236689 | 255911 |
| Time Spy | 46873 | 42614 |
| Wild Life Extreme | 108751 | 110070 |
API
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| CUDA | 12.0 | 12 |
| DLSS | DLSS 4 | DLSS 4 |
| DirectX | 12.2 | 12.2 |
| DisplayPort | 2.1b | 2.1b |
| OpenCL | 3.0 | 3.0 |
| OpenGL | 4.6 | 4.6 |
| Ray Tracing | Yes | Yes |
| Vulkan | 1.4 | 1.4 |
Games
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| 1080p High | 299 | 284 |
| 1080p Ultra | 254 | 242 |
| 1440p Ultra | 207 | 196 |
| 4K Ultra | 125 | 118 |
| CoD: Modern Warfare III | 334322260184 | 305296252176 |
| Counter-Strike 2 | 36326520299 | 36226320393 |
| Cyberpunk 2077 | 26123316077 | 25322116071 |
| Far Cry 6 | 237202190121 | 221198185112 |
| Forza Horizon 5 | 336247233187 | 309235215175 |
| Ghost of Tsushima | 221185175113 | 206170158107 |
| Hogwarts Legacy | 24919916697 | 23518915089 |
| Margin of Error | Low | High |
| Shadow of the Tomb Raider | 329302238128 | 319298224121 |
| The Witcher 3 | 364327237123 | 343311216116 |
GeekBench 6 ML
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Backend | DirectML | DirectML |
| Depth Estimation (Q) | 85572 | 90853 |
| Face Detection (HP) | 114685 | 103224 |
| Framework | ONNX | ONNX |
| Image Classification (SP) | 18948 | 16903 |
| Image Segmentation (HP) | 55109 | 48518 |
| Image Super Resolution (Q) | 58478 | 54618 |
| Machine Translation (HP) | 8184 | 7291 |
| Object Detection (SP) | 27802 | 26498 |
| Pose Estimation (Q) | 344729 | 338587 |
| Style Transfer (SP) | 687080 | 635289 |
| Text Classification (SP) | 4464 | 3894 |
GeekBench 6 OpenCL
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| API | OpenCL | OpenCL |
| Background Blur | 333 img/sec | 366.1 img/sec |
| Edge Detection | 33.7 Gpixels/sec | 35.2 Gpixels/sec |
| Face Detection | 298.5 img/sec | 248 img/sec |
| Feature Matching | 2.53 Gpixels/sec | 2.47 Gpixels/sec |
| Gaussian Blur | 34.5 Gpixels/sec | 45 Gpixels/sec |
| Horizon Detection | 21.7 Gpixels/sec | 23.3 Gpixels/sec |
| Particle Physics | 59146.9 FPS | 42574.6 FPS |
| Stereo Matching | 2850 Gpixels/sec | 3040 Gpixels/sec |
General
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Architecture | Blackwell 2.0 | Blackwell 2.0 |
| Build | Discrete | Discrete |
| Case | Desktop | Desktop |
| GPU Codename | GB202-300 | GB202-300-A1 |
| Launch price (MSRP) | $1999 | $8559 |
| Purpose | Gaming | Professional |
| Recommended CPU | -Intel Core Ultra 9 285Kor above | -AMD Ryzen 9 9955HX3Dor above |
| Released | January 7, 2025 | March 18, 2026 |
| Segment | High-end | High-end |
| Vendor | Nvidia | Nvidia |
Graphics Processing Unit
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Base Clock | 2017 MHz | 1590 MHz |
| Boost Clock | 2407 MHz | 2617 MHz |
| Compute Units (Pipelines) | 170 | 188 |
| Instructions Per Cycle | 2 IPC | 2 IPC |
| L1 Cache | 128KB per cluster | 128KB per cluster |
| L2 Cache | 96MB shared | 128MB shared |
| Ray-tracing Cores | 170 | 188 |
| Render Output Units (ROPs) | 176 | 190 |
| Shading Units | 21760 | 24064 |
| Tensor Cores | 680 | 752 |
| Texture Mapping Units (TMUs) | 680 | 752 |
Memory
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Bus | 512-bit | 512-bit |
| ECC | No | Yes |
| Effective Memory Speed | 28000 Mbps | 28000 Mbps |
| Memory Clock | 1750 MHz | 1750 MHz |
| Memory Size | 32 GB | 96 GB |
| Memory Type | GDDR7 | GDDR7 |
Passmark Graphics
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| DirectX 11 | 341 FPS | 339 FPS |
| DirectX 12 | 179 FPS | 154 FPS |
| G2D Mark | 1410 | 1391 |
| GPU Compute | 24360 Ops/s | 26607 Ops/s |
Physical
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Die Size | 750 mm² | 750 mm² |
| Fabrication Process | 4 nm | 5 nm |
| Interface | PCIe 5.0 x16 | PCIe 5.0 x16 |
| Manufacturing | TSMC | TSMC |
| TGP | 575 W | 600 W |
| Transistor Count | 92.2 billion | 92.2 billion |
| Transistor Density | 122.93 MTr/mm² | 122.93 MTr/mm² |
Raw Performance
| Specification | GeForce RTX 5090 | Nvidia RTX Pro 6000 Blackwell |
|---|---|---|
| Pixel Fill Rate | 424 GPixel/s | 497 GPixel/s |
| Texture Fill Rate | 1637 GTexel/s | 1968 GTexel/s |
Looking for More Comparisons?
Browse all available head-to-head comparisons or select new devices in the comparison hub.