The Nvidia RTX 4070 Super represents a significant improvement over its predecessor, the RTX 4070, particularly in terms of performance. Despite retaining the same 192-bit memory bus and 12GB of GDDG6X memory, the increase in CUDA cores brings its performance closer to the RTX 4070 Ti, marking a notable generational leap. This is especially evident in its enhanced 1440p and 4K capabilities.
Read our extensive review by Mario Vasquez here on our sister site, BabelTechReviews.
| RTX 4070 | RTX 4070 Super | Percentage Change | |
| GPU | AD104 | AD104 | |
| CUDA Cores | 5,888 | 7,168 | 22% |
| Ray-Tracing Cores | 46 | 56 | 22% |
| TMU | 184 | 228 | 24% |
| Base Clock (MHz) | 1,920 | 1,980 | 3% |
| Boost Clock (MHz) | 2,475 | 2,475 | 0% |
| Memory (GB) | 12 | 12 | 0% |
| Memory type | GDDR6X | GDDR6X | |
| FP32 TFLOPS | 29.15 | 35.48 | 22% |
| Memory Bus | 192-bit | 192-bit | 0% |
| Bandwidth (GB/s) | 504.2 | 504.2 | 0% |
| TDP (W) | 200 | 220 | 10% |
| MSRP at Launch | $599 | $599 | 0% |
| Street Price (January 2024) | $534 | $599 | 12% |
Graphics card comparison of RTX 4070 vs. RTX 4070 Super.