Today we're finally publishing our GeForce RTX 4060 Ti 16 GB review. While we usually receive samples from NVIDIA or the various board partners to prepare our coverage leading up to the official launch, things were different this time. NVIDIA isn't producing a Founders Edition of the 16 GB RTX 4060 Ti, and they couldn't help us with a board partner sample either. When asked, all the board partners declined to send a sample as well—an unprecedented situation. The GeForce RTX 4060 Ti 16 GB was announced together with the rest of the RTX 4060 family, on May 18th. The three models that make up the RTX 4060 Series are the RTX 4060 non-Ti 8 GB, priced at an MSRP of $300 (I've tested 10 models), the RTX 4060 Ti 8 GB, priced at an MSRP of $400 (8 reviews) and the RTX 4060 Ti 16 GB, which is covered in this review. The 16 GB version comes at whopping $500, which brings it fairly close to the $600 price point of the RTX 4070 non-Ti.

Just like the GeForce RTX 4060 Ti 8 GB, the 16 GB version is based on the NVIDIA AD106 graphics processor, which also powers several laptop GPU models—for desktop it's the first and only release so far. RTX 4070 non-Ti and RTX 4070 Ti are both based on AD104, RTX 4060 non-Ti uses AD107. As expected, all RTX 4060 series cards are built using the NVIDIA Ada architecture, which not only brings improvements to efficiency and ray tracing, but also comes with the DLSS 3 Frame Generation feature, which is a game changer, especially in the lower-end segments, where reaching decent FPS matters more than anything else.

In the past, NVIDIA increased the core counts for some models with extra memory, to give them an extra boost in scenarios where the extra memory can't make a difference. There's no such thing on the RTX 4060 Ti 16 GB. All specs are identical to the 8 GB model, 4352 cores, 2535 MHz Boost, 2250 MHz GDDR6 memory. The bus width hasn't changed either, NVIDIA uses a clamshell configuration for memory, in which two memory chips share a 32-bit link to the memory controller. This means that memory bandwidth is exactly the same as on the 8 GB card. The only difference, besides the VRAM capacity increase to 16 GB, is that the power limit has been raised to 165 W, a 5 W bump, to account for the higher power draw of the memory subsystem.

In this review we're testing the Gainward RTX 4060 Ti 16 GB Panther OC, which is a factory overclocked custom design running at a rated clock frequency of 2595 MHz, or +60 MHz over NVIDIA reference (+2.4%). Right now it's impossible to find a card in stock that runs at stock clocks, just underclocking the card will not work correctly to simulate a reference card, because NVIDIA's Boost algorithm doesn't work in such a linear fashion. I've reviewed dozens of cards with small factory overclocks like this and the gains are typically in the 1% range, very small.

Averaged over the 25 games in our test suite, the RTX 4060 Ti 16 GB ends up a whole 0% faster at 1080p Full HD, 1% at 1440p, and 2% at 4K. While there's games that do show small performance improvements, the vast majority of titles runs within margin of error to the RTX 4060 Ti 8 GB Founders Edition. The best gains with 1080p are seen in Plague Tale (+2%), Deathloop (+1.6%), F1 23 (+1.2%), Far Cry 6 (+0.9%) and Cyberpunk (+0.8%). At higher resolution these gains do improve, but only marginally. The only exception is The Last of Us, which gains +19% at 1440p and +21% at 4K. F1 23 gains 10% at 4K, too, but RTX 4060 Ti is fundamentally not a card for 4K in modern AAA titles. The Last of Us 4K runs at 27 FPS with the 16 GB card and 22 FPS with the 8 GB model—both too slow for a serious gaming experience—and that's the title with the biggest gains.

The differences are a little bit more pronounced when looking at Minimum FPS: 1080p: +1%, 1440p: +2%, 4K: +6%. I'm still not convinced. No doubt, you can find plenty of evidence for 16 GB superiority if you make it your mission to find situations where an 8 GB card runs into trouble, and a 16 GB card does not. But for the vast majority of games out there, 99% at least, there is no difference between a 8 GB and 16 GB card, at 1080p and 1440p. Even in those rare cases, justifying the price increase of $100 is impossible in my opinion. The better approach is to dial down a setting or two, to reduce VRAM pressure. I think that's not unreasonable for a x60 card, even though it might be if you consider the $400/$500 price point of these cards.

As expected, ray tracing performance of RTX 4060 Ti is clearly better than its AMD counterparts. With RT enabled, the RTX 4060 Ti matches the Radeon RX 6800 XT, which is positioned roughly two tiers above it. AMD's Radeon RX 6700 XT is a whopping 30% slower. Still, I'm not sure if ray tracing really matters in this segment. The technology comes with a big performance hit that I find difficult to justify, especially when you're already fighting to stay above 60 FPS in heated battles. The extra VRAM does help a little bit more in ray tracing titles, but here, too, nothing worth losing sleep over.

Probably the most important selling point for the GeForce RTX 4060 Ti is support for DLSS 3 Frame Generation. The algorithm takes two frames, measures how things have moved in those two frames and calculates an intermediate frame in which these things moved only half the distance. While this approach is definitely not problem-free, especially when pixel-peeping at stills or slowed down video, in real-time it's nearly impossible to notice any difference. As you run at higher FPS and resolution it becomes even more difficult, because the deltas between each frame are getting smaller and smaller. Being able to double your FPS is a huge capability, because it means you can enable ray tracing for free, or game at higher resolutions. Of course you are limited to games with DLSS 3 support, of which there are currently around 40, mostly AAA titles, but not every title will support it. AMD doesn't have anything similar, they announced that FSR 3 exists last year and since then we haven't seen a single demo, with no updates at Computex either.

Gainward did a good job with their cooler, even though it does look a little bit plasticky, like a $199 card from not that long ago. There's sufficient cooling potential to keep the card at good temperatures and low noise levels of 32 dBA. While that is a little bit louder than the 8 GB Founders Edition cooler, the subjective differences are minimal. The extra 8 GB of VRAM really don't put out much heat to affect the overall balance between heat and noise, but since the memory chips are located on the back side of the PCB, that part does get a little bit hotter. The metal backplate with thermal pads definitely helps to distribute the heat and while it gets fairly warm, I wouldn't call it "hot."

As expected, the energy efficiency of the RTX 4060 Ti 16 GB is a little bit worse than the 8 GB variant—it's still excellent. The additional four chips increase power draw by 6 to 10 W, which is pretty much negligible in gaming, but in non-gaming states it is more noticeable. As mentioned before, NVIDIA increased the default power limit by 5 W, to ensure the GPU has the same amount of power available as on the 8 GB card. Gainward was very generous with the manual adjustment range for power limit and gives us up to 200 W, which is more than on nearly all RTX 4060 Ti models.

There's nothing to report for overclocking, it works just like on the 8 GB variant, the extra VRAM doesn't affect memory overclocking, or overclocked FPS. We gained 9.1% in real-life performance which is quite decent, and much more significant than what the extra VRAM provided. I did a quick test with power limits increased to maximum, but couldn't get any additional performance gains, which confirms that power limits are not an issue on the RTX 4060 Ti 16 GB.

So far the RTX 4060 Ti 16 GB has turned out to be a big nothingburger. Still, I can definitely feel the "future-proofing" argument, which is not completely unreasonable. No doubt, future games will come with increased VRAM requirements. Personally I feel like these requirements will increase very slowly, since they've now pretty much matched the current-gen consoles. Unless there's a push for more memory on the consoles, I doubt developers will give premium treatment to the PC platform—ports are actually getting worse each year. That same argument does work in favor of 16 GB though. When developers fail at their primary job (writing proper code), then you get ports like The Last of Us, which needs a ton of VRAM for no apparent reason. The Steam ratings explode, the publisher puts pressure on the developer and things usually get fixed, but that takes time. With more VRAM on your graphics card you might be able to ride out the launch drama. However, no amount of VRAM will fix stuttering from shader compilation.

If the price increase for the extra 8 GB was $20 I would definitely consider it, even though gains today are negligible. But $100 is simply too much. There's no way this can be justified, especially when current GDDR6 memory pricing suggests a ~$10 cost increase for the card makers. At $500 the RTX 4060 Ti 16 GB is extremely bad value, you'll be much better off buying a RTX 3080 or Radeon RX 6800 XT for $500. The highly discounted RX 6800 non-XT for $430 will give you 16 GB VRAM, too, for $70 less, with better raster performance but slightly lower RT perf. Even GeForce RTX 4070 offers a better price/performance ratio than the RTX 4060 Ti 16 GB. While it costs $600 (+20%), it offers 30% higher performance at Full HD, even more at 1440p and 4K. If you're willing to look at used last-generation cards there's plenty of interesting options here, too, but those are less efficient and lack support for NVIDIA's DLSS 3 feature, which is extremely useful to boost frame rates.

I do wonder why NVIDIA had so little love to share for this 16 GB version. It would have been easy for them to enable more shaders, use faster memory, or even switch to GDDR6X memory. I guess the strategy here is that a significant percentage of their customer base will think "16 GB is twice as much as 8 GB," and that "+$100 for future-proofing" could be a reasonable upgrade over the base RTX 4060 Ti, without spending +$200 for the RTX 4070. The fact is that NVIDIA owns the GPU market and AMD seems happy with their small market share, selling similar products at 10% less than NV. Maybe the upcoming Radeon RX 7800 XT and RX 7700 XT can change that.

I bought the tested card in retail, do consider supporting us on Patreon if you appreciated this coverage.