Nvidia GeForce RTX 5060 Ti 16GB review: decent gen-on-gen uplifts, but RTX 5070 offers better value

The RTX 5060 Ti is here, and with it, the first ostensibly “mainstream” graphics card from Nvidia’s 50-series Blackwell architecture – though with MSRPs at $379/£349 for the 8GB card and $429/£399 for the otherwise identical 16GB variant, there’s still an appreciable gap over the entry-level RTX 5060 at $299.

Having both 8GB and 16GB models available muddies the waters somewhat, so let’s make things simpler: don’t consider buying the 8GB model. Reporting from Digital Foundry and the tech press at large has made it increasingly clear that 8GB isn’t sufficient for anything other than the most entry-level cards, so in a world where you’ll likely be keeping your GPU for three to five years, it’s best to make alternative arrangements. The fact that Nvidia isn’t sending out samples of the 8GB card speaks for itself, really.

With the 16GB card in hand then, we’d ideally like to see a card that offers a touch more value than the RTX 5070, while at the same time delivering a palpable boost over the lacklustre RTX 4060 Ti – which also came in two guises: a disappointing 8GB model and a vastly overpriced 16GB alternative. This time around, the 8GB version is cheaper, while the price premium for the 16GB version is much reduced. It’s a welcome move.

1 of 5 Caption Attribution

Let’s look at specs. With 4608 CUDA cores, we’re only seeing a six percent increase over the RTX 4060 Ti, with the 5070 delivering a 33 per cent increase – albeit with slightly slower boost clocks. The 5060 Ti does has a huge memory bandwidth advantage over its predecessor, to the tune of 55.5 percent, but again, the 5070 is far ahead – a straight 50 percent increase. So in theory then, the 5060 Ti shouldn’t trouble the 5070 at all. However, in terms of gen-on-gen increases, we should see some impressive results – depending on how much the Blackwell architecture taps into that bandwidth – and of course the extra 20W ceiling on total board power.