The Nvidia GeForce RTX 5070 was solely absolutely revealed this week at CES 2025, however already the GPU is inflicting fairly a stir. Though arguably the least highly effective of the simply unveiled 50-series, the RTX 5070 has been the topic of numerous headline-worthy claims. Alas, most people must wait till the 50-series launches this February to see if it is definitely worth the hype—however tech journalists aren’t most people.
Case in level, PCGamesN has already seen the RTX 5070 in motion and has since shared its findings. So, what is the reality behind Nvidia’s CEO Jen-Hsun Huang on-stage declare that the RTX 5070 will supply “RTX 4090 efficiency at $549”? Turns out, there’s rather a lot to that—however not with out caveats.
First, the excellent news: whereas taking part in Marvel Rivals, PCGamesN journalist Ben Hardwidge discovered that the RTX 5070 would not simply supply comparable efficiency, however truly eclipses the RTX 4090, what was as soon as Nvidia’s flagship card. However, the not-really-bad-news, however extra I’m-not-sure-how-to-feel-about-this-news is that we’ve got the GPU’s new Tensor cores and their assist of Nvidia’s simply introduced DLSS 4 Multi Frame Generation to thank.
It’s hardly a shocking transfer. Nvidia has been calling time on raster graphics rendering for years, with Huang himself memorably saying again in September 2024, “We cannot do laptop graphics anymore with out synthetic intelligence. We compute one pixel, we infer the opposite 32. I imply, it is unbelievable.”
To be clear although, the identify of DLSS 4’s recreation is most impressively body technology. With regards to the RTX 5070, DLSS 4 makes use of AI and in-game information to generate as much as three full frames between extra historically rendered frames. To put it somewhat crudely, DLSS 4 is what’s providing you with that further bang on your buck.
At CES 2025, Nvidia arrange two machines for commerce present friends to check—one kitted out with the RTX 5070 and multi-frame technology, and one utilizing the older RTX 4090 with commonplace body technology. Besides that main distinction, each machines had been operating Marvel Rivals at 4K with the very best quantity of DLSS supported by their respective graphics card. The end result? Hardwridge writes that the RTX 5070 constantly outperformed the older card, outputting within the area of 240 fps in comparison with the RTX 4090’s common of 180 fps.
All of that stated key tech specs, corresponding to which CPU was used for these comparability builds, stay unclear, so it is exhausting to say conclusively simply how replicable these numbers are. It additionally maybe goes with out saying however the older card would not assist DLSS 4’s multi body gen, so one may quibble about how invaluable a comparability this really is.
Furthermore, even Nvidia pressured that this comparability level is a little bit of an outlier, as Marvel Rivals seems to be notably well-optimised for the RTX 5070; when examined on titles that aren’t Marvel Rivals, there’s apparently a a lot smaller efficiency hole between the 2 playing cards.
Still, that means efficiency that after value over a grand and a half for the RTX 4090 is now inside attain for a lot much less. Provided DLSS 4 is concerned, in fact.
