NVIDIA RTX 4080 12GB just 12% Faster than the RTX 3080, Slower than the 3090 Ti in Every Game

NVIDIA has released the first official performance figures of its RTX 4080 lineup. Unlike previous generations, the x80 Ada family is composed of two very different SKUs, misleadingly distinguished using their memory buffers. The 12GB variant is based on the AD104 die, which we all thought would power the RTX 4070, while the 16GB model features the AD103 die.

The difference in memory buffers won’t affect real-world gaming performance by much. However, the differing core counts, bus widths, and the cache will nudge the 12GB SKU closer to the RTX 3080, essentially making it the successor to the 3070. NVIDIA’s official benchmarks show the RTX 3090 Ti beating the RTX 4080 12GB at native in all three titles. (valium) And you can be sure that these are cherry-picked examples showing the average, if not peak, raster gains.

The deltas between the two variants of the RTX 4080 are also rather glaring. The 16GB model is 23-28% faster than its twin in most cases, with DLSS 3 bringing it closer to 20%. There’s also a sizable deficit between the capabilities of the RTX 4090 and 4080 16GB. In some cases like Plague Tale Requiem, the former is nearly 2x faster at native, and 37% with DLSS 3 enabled.

We’ve never seen such a large deficit between the x80 and x90 offerings of a given GeForce family. With a price tag of $1,199, this makes the RTX 4080 16GB look like an even worse deal than its 12GB sibling. At $899, the cheaper variant offers a marginal improvement over the $699 RTX 3080, while the other barely manages to beat the now roughly $1,000 RTX 3090 Ti. These figures are derived from the frame rates at native resolution without DLSS 3 but let’s face it. You don’t wanna shell out a grand for a couple dozen supported titles.