NVIDIA GeForce RTX 5060 Ti 16GB Review

Lower price than last gen, same amount of VRAM

The GeForce RTX 5060 family of cards is about to be released, and we have a 16GB version of the RTX 5060 Ti in for review. NVIDIA partners will be offering both 16GB and 8GB variants, and the fact that any card outside of the budget realm is being offered at all with just 8GB of VRAM is something that only NVIDIA can explain. Games haven’t suddenly become less demanding, and 8GB was demonstrably inadequate (not to mention wildly unpopular) in 2023.

The pricing strategy from NVIDIA for this launch does indicate that the company is at least aware of negative feelings about 8GB cards, as $20 has been shaved from the MSRP of the RTX 5060 Ti 8GB, making it start at $379 USD (theoretically), instead of the RTX 4060 Ti 8GB‘s $399 launch MSRP almost two years ago. A bigger drop comes from the 16GB version of the RTX 5060 Ti, which is (theoretically) offered for $429, down $70 from the $499 (theoretical) starting price of the RTX 4060 Ti 16GB.

Of course, pricing is all subject to change, and availability is always a painful subject these days. Really, it’s been such a long time since a graphics card review did not have to disclaim any listed MSRP or bemoan availability that I’ve largely forgotten what it was like when cards like the GTX 1070 were released at a good price, and could actually be found for that amount (not to mention that 8GB was a lot more attractive in 2016).

Here is a look at specs for the new RTX 5060 series, compared to the previous RTX 4060 series:

The move to 28 Gbps GDDR7 might be the most interesting thing about the hardware with this launch, as it is a massive upgrade from the RTX 4060 family’s 17 – 18 Gbps GDDR6. The memory bandwidth increase alone – from 288 GB/s to 448 GB/s – could make a significant difference, depending on the game.

Beyond GDDR7 and DLSS 4 support, there isn’t a lot to talk about (outside of exciting things like encode/decode performance) compared to the RTX 4060 Ti launch. We are still – nearly two years later – looking at the reality of an 8GB RTX 5060 Ti card. Granted, we were not sampled an 8GB variant like we were at the RTX 4060 Ti launch in 2023, but their very existence really seems to upset a lot of people.

Speaking of samples, we received a PNY RTX 5060 Ti OC Dual Fan model, part number VCG5060T16DFXPB1-O, linked here (PDF).

Some Performance Numbers

As we embarked upon the task of not only testing this new graphics card, but also some logical comparisons, the reality of the 2-year gap between these generations was evident. Back in May of 2023 we were still using an Intel Core i5-13600K system for “mainstream” testing, and as we have since brought all testing to the same AMD Ryzen 7 9800X3D platform, the comparisons presented here were either done in a hurry, or re-used from recent high-end GPU reviews. As such, the settings used – 2560×1440 at the highest preset – may be unrealistic for the sub-$500 cards on the charts.

The RTX 5060 Ti 16GB was 21% faster than the original RTX 4060 Ti in Steel Nomad, and 27% faster in Speed Way. In both cases the original RTX 4070 is faster. AMD’s RX 7800 XT does as well or better overall in these 3DMark tests, making it one of the best sub-$500 cards that you can’t buy at that price.

The RTX 5060 Ti 16GB was ~39% faster than the original RTX 4060 Ti in Cyberpunk 2077, and actually beat the RTX 4070 in that test. In The Talos Principle 2 the RTX 5060 Ti 16GB was ~26% faster than the RTX 4060 Ti, but only ~11% faster in the old Metro Exodus Enhanced Edition test.

There should be more tests here, but a reboot of the “mainstream” testbed – and some 1080/Ultra results (along with additional games) should be coming in the next review. As it stands, the results above were achieved using settings that push the limits of the imaginary sub-$500 segment, with cards in the theoretical $549 to $749 segment really more appropriate for these settings.

Closing Thoughts

The addition of GDDR7 for all members of the RTX 5060 family really makes a difference, at least based on the results from our RTX 5060 Ti 16GB sample. If NVIDIA had released the cards with GDDR6, the only significant difference compared to the RTX 4060 family would probably be DLSS 4, and of course the Frame Generation that seems to be the focus of NVIDIA’s consumer graphics business now.

If you look at NVIDIA’s official performance charts, each is presented using the “max frame gen level supported by each game/GPU”, so the era of hardware FPS appears to be over now. So-called “fake frames” are now the standard, and we will have to re-test using some DLSS + FG by the time the RTX 5060 (non-Ti) is here. As it stands, these cards do seem to be well positioned as great 1080p/Ultra options, and when DLSS is enabled that resolution can easily be pushed into enthusiast territory.

But all of this is academic if the GPUs aren’t available anywhere near their MSRP, or just not available at all post-launch. We will see if anything has changed since the RX 9070 Series launch. Don’t get your hopes up – unless you live near a Micro Center, of course.