Quick/Comparisons
34 | Gaming | 16 |
41 | Workstation | 25 |
31 | AI/ML |
Compare with something else?
Compare GeForce RTX 3070 with another →22 |
43 | Energy Efficiency | 100 |
37 | Quick Comparison Final Score | 33 |
| GA104 (GN20-E5) | Graphics Processor | Panther Lake |
| 5120 | Cores | 1536 |
| 160 / 80 | TMUs / ROPs | 48 / 24 |
| 2998.92 | Blender GPU | 1268.79 |
| 79 | Forza Horizon 5 | 38 |
| 114 | The Witcher 3 | 55 |
| 150 | Counter-Strike 2 | 60 |
| 94 | Far Cry 6 | 49 |
| 61 | Hogwarts Legacy | 21 |
| 88 | Call of Duty | 32 |
| 56 | Ghost of Tsushima | 25 |
| 73 | Cyberpunk 2077 | 33 |
| 116 | Tomb Raider | 52 |
| 92 | Average FPS | 41 |
| 114 | 1080p High | 52 |
| 92 | 1080p Ultra | 41 |
| 68 | 1440p Ultra | 29 |
| 38 | 4K Ultra | 14 |
| Low | Margin of Error | Low |
| 118499 | GB6 Compute Score | 64546 |
| 170.3 img/sec | Background Blur | — |
| 113.6 img/sec | Face Detection | — |
| 5.59 Gpixels/sec | Horizon Detection | — |
| 7.94 Gpixels/sec | Edge Detection | — |
| 6.19 Gpixels/sec | Gaussian Blur | — |
| 1.16 Gpixels/sec | Feature Matching | — |
| 467 Gpixels/sec | Stereo Matching | — |
| 15131.3 FPS | Particle Physics | — |
| OpenCL | API | — |
| 15105 | G3D Mark Score | 9029 |
| 630 | G2D Mark | 786 |
| 137 FPS | DirectX 11 | 70 FPS |
| 63 FPS | DirectX 12 | 50 FPS |
| 6702 Ops/s | GPU Compute | 3702 Ops/s |
| 17250 | GB6 ML Single Precision | 11871 |
| 27936 | GB6 ML Half Precision | 20120 |
| 11249 | GB6 ML Quantized | 7879 |
| 6969 | Image Classification (SP) | 5487 |
| 17142 | Image Segmentation (HP) | 16864 |
| 16426 | Image Super Resolution (Q) | 8788 |
| 31628 | Face Detection (HP) | 25029 |
| 74413 | Pose Estimation (Q) | 18852 |
| 2522 | Text Classification (SP) | 2115 |
| 3639 | Machine Translation (HP) | 4266 |
| 9002 | Object Detection (SP) | 7449 |
| 27790 | Depth Estimation (Q) | 22207 |
| 160005 | Style Transfer (SP) | 95039 |
| ONNX | Framework | ONNX |
| DirectML | Backend | DirectML |
| 103-130 GPixel/s | Pixel Fill Rate | 60 GPixel/s |
| 206-259 GTexel/s | Texture Fill Rate | 120 GTexel/s |
| 13.2-16.6 TFLOPS | FLOPS (FP32) | 7.7 TFLOPS |
| GDDR6 | Memory Type | System Shared |
| 8 GB | Memory Size | — |
| 1500-1750 MHz | Memory Clock | — |
| 12000-14000 Mbps | Effective Memory Speed | — |
| 256-bit | Bus | — |
| No | ECC | No |
| 448 GB/s | Memory Bandwidth | — |
| PCIe 4.0 x16 | Interface | PCIe 5.0 |
| 80-150 W (configurable) | TGP | — |
| Samsung | Manufacturing | TSMC |
| 8 nm | Fabrication Process | 3 nm |
| 392 mm² | Die Size | — |
| 17.4 billion | Transistor Count | — |
| 44.39 MTr/mm² | Transistor Density | — |
| 12 | DirectX | 12.1 |
| 1.2 | Vulkan | 1.4 |
| 4.6 | OpenGL | 4.6 |
| 2 | OpenCL | 3 |
| 8.6 | CUDA | — |
| Yes | Ray Tracing | Yes |
| DLSS 2 | DLSS | No |
| 1.4a | DisplayPort | 2.1 |
| Nvidia | Vendor | Intel |
| Discrete | Build | Integrated |
| January 12, 2021 | Released | January 5, 2026 |
| Laptop | Case | Laptop |
| Gaming | Purpose | Gaming |
| High-end | Segment | Mid-range |
| Ampere | Architecture | Celestial |
| GA104 (GN20-E5) | GPU Codename | Panther Lake |
| -Intel Core i7 12700Hor above | Recommended CPU | -Intel Core Ultra X9 388Hor above |
| 10214 | Steel Nomad Lite Score | 4242 |
| 10477 | Time Spy | 6595 |
| 45215 | Solar Bay | 26399 |
| 6247 | Port Royal | 3886 |
| 26729 | Fire Strike | 16770 |
| 21137 | Wild Life Extreme | 11707 |
| 82533 | Night Raid | 68498 |
| 780-1215 MHz (configurable) | Base Clock | 300 MHz |
| 1290-1620 MHz (configurable) | Boost Clock | 2500 MHz |
| 5120 | Shading Units | 1536 |
| 160 | Texture Mapping Units (TMUs) | 48 |
| 80 | Render Output Units (ROPs) | 24 |
| 40 | Compute Units (Pipelines) | 12 |
| 160 | Tensor Cores | — |
| 40 | Ray-tracing Cores | — |
| 128KB per cluster | L1 Cache | 64KB per cluster |
| 4MB shared | L2 Cache | 16MB shared |
| 2 IPC | Instructions Per Cycle | 2 IPC |