True to the leaks, NVIDIA has unveiled the standard GeForce RTX 4070 — and it might hit the sweet spot, provided you're willing to accept the price hikes that have defined the RTX 40 lineup. The $599 desktop GPU is considered a successor to the RTX 3070 Ti and offers a similar 5,888 CUDA cores, but adds more RAM (12GB instead of 8GB), higher clock speeds (1.92GHz base versus 1.58GHz), more level 2 cache (36MB compared to 4MB) and greater floating point computing power (29 teraflops versus 22). At the same time, it reduces the typical gaming power consumption from 290W to 200W. Some games quickly ran into the 8GB ceiling of RTX 3070 variants, so the extra memory alone may have a significant effect.
NVIDIA pitches the RTX 4070 as ideal for 1440p gaming at over 100 frames per second with ray tracing turned on, although you'll have to enable DLSS3 upscaling to hit that target. While many games don't support the feature yet (over 50 do so far), this might make it feasible to play big titles like Cyberpunk 2077 and Spider-Man: Miles Morales at maximum visual detail without compromising on frame rates.
The RTX 4070 will be available tomorrow, April 13th, and will include a Founders Edition card from NVIDIA itself. In other words, the $599 price is more likely to reflect what you can actually buy. You might not have to pay extra for a vendor's factory-overclocked board or otherwise accept artificial markups.
Whether or not the new GPU is a good deal depends on your perspective. NVIDIA sees the RTX 4070 as an upgrade path for users coming from a GTX 1080 or RTX 2070. You're paying similar money for a major leap in performance. It's also decidedly more affordable than the $799 RTX 4070 Ti. However, the new model still costs $100 more than the base 3070. Like it or not, the days of sub-$500 upper mid-range cards appear to be over in NVIDIA's world.
This article originally appeared on Engadget at https://www.engadget.com/nvidias-geforce-rtx-4070-comes-with-12gb-of-ram-for-599-130002404.html?src=rss