Our coverage of the GeForce RTX 4070 Ti Super includes ten cards: ASUS RTX 4070 Ti Super STRIX, ASUS RTX 4070 Ti Super TUF, Colorful RTX 4070 Ti Super Vulcan W, Gainward RTX 4070 Tio Super Phoenix GS, Galax RTX 4070 Ti Super EX White, Gigabyte RTX 4070 Ti Gaming OC, MSI RTX 4070 Ti Super Ventus 3X, Palit RTX 4070 Ti Super JetStream OC, PNY RTX 4070 Ti Super Verto and Zotac RTX 4070 Ti Super Trinity.

With the GeForce RTX 40 Super series, NVIDIA is refreshing their Ada-based lineup for 2024. At CES Las Vegas, the company announced three new SKUs: GeForce RTX 4070 Super, RTX 4070 Super Ti and RTX 4080 Super. While we reviewed RTX 4070 Super last week, today, the review embargo for RTX 4070 Ti Super has lifted, the custom design card reviews go live tomorrow, and the RTX 4080 Super will be reviewed later this month. While RTX 4070 Super is built using the same AD104 GPU as the RTX 4070, the RTX 4070 Ti Super is different. RTX 4070 Ti Super uses the bigger AD103 GPU that powers RTX 4080. The underlying reason is that RTX 4070 Ti non-Super already maxes out the GPU core count of the AD104 GPU. Additionally, AD104 cannot support a 16 GB VRAM configuration—only 12 GB and 24 GB. With the GPU upgrade you also get an increase in GPU cores, to 8448, which is a 10% increase over the 7680 cores in RTX 4070 Ti. As expected for a 16 GB VRAM configuration, the VRAM bus interface is now 256-bit wide, which is a significant boost over the 192-bit bus on the RTX 4070 Ti. The L2 cache size is unchanged—still 48 MB, or 16 MB less than RTX 4080. In terms of ROP count you're now getting 96 (this is the correct number, I confirmed with NVIDIA, 112 is a typo in the reviewer's guide), which is a 20% increase over the 80 ROPs on 4070 Ti. The RTX 4080 is still expected to be significantly faster, mostly thanks to its 9728 GPU cores, larger cache and 112 ROPs.

The Gigabyte RTX 4070 Ti Super Gaming OC is the company's middle model for the RTX 4070 Ti Super. You get a triple-slot, triple-fan cooling solution and factory overclock. Compared to the NVIDIA baseline frequency of 2610 MHz boost, the OC is rated for 2655 MHz, or +1.7%. Averaged over the 25 games in our test suite, at 1440p, we find the card only 7% faster than 4070 Ti non-Super, which is much less than expected. At 4K, the gains are slightly bigger with 10%, but it's definitely not the 15%+ jump that we got with the RTX 4070 Super. Still, these gains help make up ground against AMD's RX 7900 XT, which still remains 2% faster in a pure raster scenario. Compared to last generation's flagship, the RTX 3090 Ti, the new Super card can beat it by a small margin of 5%, which is still impressive. You're basically getting last gen's x90 Ti performance with the x70 Ti Super. Compared to RTX 4070 Super, the performance uplift is 14% and the new card is 33% faster than the RTX 4070 non-Ti non-Super, too. The fastest card from Team Red, the Radeon RX 7900 XTX, is still 16% ahead, 20% at 4K—RTX 4080 Super is designed to rival this card.

With these performance numbers RTX 4070 Ti Super is a perfect match for 1440p with maximum settings, it's actually slightly overkill, which means that the card is a decent option for 4K monitors, too, or for 1440p at 120/144 Hz. While you won't be able to game at 4K60 at highest settings, just dropping them down a bit should help get those 60 frames and there's always the various upscaling technologies, especially if you plan on enabling ray tracing. Just like the other GeForce 40 cards, RTX 4070 Ti Super has support for all of NVIDIA's DLSS technologies: NVIDIA DLSS 2 upscaling, DLSS 3 frame generation and DLSS 3.5 ray reconstruction. On top of that you can enable AMD FSR 2 and FSR 3 in games, because those technologies work on all GPUs from all vendors. Basically this means that you'll be covered in terms of upscaling and frame generation. While DLSS 3 is definitely the leading solution right now, with the best game support, AMD is pushing hard and their frame generation solution will come to several major titles in 2024. From a technology perspective, DLSS 3 is superior, because it uses the optical flow hardware unit in Ada GPUs, and NVIDIA Reflex will help bring down the input latency.

The biggest selling point of the RTX 4070 Ti Super vs the RTX 4070 Ti non-Super is the increased VRAM size of 16 GB. RTX 4070 Ti's 12 GB VRAM size has been a constant topic for debate on tech forums, so it makes a lot of sense that NVIDIA is giving us a 16 GB option now, and at pretty reasonable pricing, unlike RTX 4060 Ti 16 GB. Contrary to more cores or higher clocks, more VRAM will not make all games run faster automatically. Across all the 100+ game tests, (25 raster + 10 RT) x 3 resolutions, we only identified two cases where 16 GB results in a meaningful improvement over 12 GB: The Last of Us 4K and Alan Wake 2 RT at 4K. No doubt, you will be able to find more such results with other titles, too, but the vast majority of games out there will not see any meaningful improvement from the 16 GB upgrade. I'm sure that this will change in the coming years, with more and more games increasing their VRAM requirements, but I don't think that a 12 GB card will suddenly turn out to be useless in 2024 and 2025. You also have to consider that as soon as you enable upscaling, the actual render resolution is reduced, which lowers the VRAM usage significantly. Still, given all the drama about 12 GB VRAM—people can finally put their money where their mouth is and grab the RTX 4070 Ti Super 16 GB.

A secondary effect of the 16 GB VRAM capacity is that the bus width is increased from 192-bit to 256-bit (or +33%). This is required, because to achieve 16 GB, you need to install eight 2 GB memory chips, each having a 32-bit interface to the GPU. With just 12 GB and six chips a 192-bit interface is sufficient (6 x 32 =192). This 33% increase in bus width leads to an equivalent increase in memory bandwidth, which should help provide an additional performance boost. Looking at my data I'm not so convinced. While the card does have slightly better scaling than RTX 4070 Ti 12 GB, the RTX 4080 is still able to pull away at higher res. It seems that what matters more for performance scaling is the L2 cache size and not the VRAM bus width. Unfortunately NVIDIA did limit the 4070 Ti Super to 48 MB L2 cache, while the RTX 4080 gets the full 64 MB.

As expected, ray tracing works very well on the GeForce RTX 4070 Ti Super, clearly offering a superior experience than what Radeon RX 7900 XT, and often even RX 7900 XTX, can achieve. On average, the RTX 4070 Ti Super offers 22% higher FPS with RT than 7900 XT, which is quite a bit. NVIDIA's new card also shows better RT performance numbers than RX 7900 XTX in most games—if you're betting on ray tracing, then definitely opt for the RTX 4070 Ti Super. That doesn't mean that RT is unusable on AMD, it's just running considerably slower, because the dedicated RT units in their cards are weaker, so more tasks are offloaded to the GPU shaders.

With our new 2024 test suite, we are adding a section focused on testing GPU Compute, which is becoming more and more important every day. Emerging AI Technologies like ChatGPT, Stable Diffusion and others are transforming the world. While the use of compute on the consumer desktop is limited today, it will grow considerably. For our first round of testing we've picked three real-life workloads that allow us to get a feel what to expect. Here, NVIDIA is the clear leader with a substantial advantage over both AMD and Intel. It's not only about performance, but also about the software ecosystem, which is much more advanced on the NV side, but the other players are working hard to catch up, I'm sure.

Gigabyte's triple-fan, triple-slot thermal solution does a good job keeping the card cool and reaches very good temperatures. Noise levels could be better though, it seems Gigabyte focused a bit much on temperatures. With 36.6 dBA, the card is the loudest RTX 4070 Ti of the 10 cards that we've tested. The secondary "quiet" BIOS helps a bit, but doesn't have enough of an impact—it reaches 33.1 dBA, which is still louder than most cards tested. I'm not sure why Gigabyte is so conservative, especially with the secondary BIOS. If it's labeled "quiet" (actually "silent"), then people who enable it will be willing to accept an increase in temperatures for lower noise levels. Our noise-normalized apples-to-apples cooler comparison test shows that Gigabyte's cooler sits roughly in the middle of the test group, which is a bit worse than expected, especially considering the eight heatpipes, but it should still be easy to build a low-noise solution with that. Compared to the 4070 Ti non-Super Gaming OC, the cooler has been downgraded a bit—the vapor-chamber design is no longer used. As expected for a modern graphics card in 2024, the fans will stop spinning when not gaming, for the perfect noise-free experience.

Power efficiency of the RTX 4070 Ti Super is virtually identical to the RTX 4070 Ti non-Super, despite of the additional VRAM and different GPU used. With around 280 W during gaming, the RTX 4070 Ti Super is a highly efficient graphics card, considering the performance offered—any half-decent PSU will be able to power it just fine. Gigabyte's card consumes a bit more power with close to 300 W, but that's expected, given the factory OC. While there is still some controversy around the 12-pin power connector, all RTX 4070 Ti Super cards come with it, and I'm a big fan, even though I'm not sure if the location in the middle of the card is the best possible choice.

There is not much to report on overclocking, other than it works, is easy to do and will yield you around 6 to 7%, which is the typical range that we've been seeing from most cards in recent years. It's a shame that most cards priced at MSRP don't allow any manual power limit increases, Gigabyte is more generous here, giving us +12%, which is better than most other cards that we've tested, up to 320 W.

NVIDIA has set a $800 MSRP for the GeForce RTX 4070 Ti Super, which matches the RTX 4070 Ti's MSRP. The RTX 4070 Ti is now end-of-life and will be sold off while supply lasts—it has been priced at $750 for many weeks. While technically NVIDIA is replacing the 4070 Ti with the Ti Super at the same price, the actual market conditions are slightly different, and the Super slots in at +$50 above the non-Super. Just like in 2023, GPUs in 2024 are not and will not be cheap. The strongest alternative for RTX 4070 Ti Super is AMD's Radeon RX 7900 XT, which had its price lowered from $760 to $710 last week, no doubt in anticipation of RTX 4070 Ti Super reviews—smart move by AMD. While I would spend the extra +$40 for RTX 4070 Ti Super any day, I'm not so sure about $90. You get slightly better raster performance from AMD and an additional 4 GB VRAM. On the other hand, AMD's card lacks DLSS 3, RT performance and draws slightly more power.

Priced at $850, the Gigabyte Gaming OC isn't that much more expensive that the MSRP models (+6.3%), but I'm still not convinced that I'd be willing to spend the extra. Especially the ASUS TUF at MSRP offers a much stronger package—better cooler, much lower noise, similar temperatures, no factory OC though—but at MSRP. Even when the ASUS TUF turns out to be $830 or $840 I'd still prefer that over the Gaming OC. If both cards are evenly matched in price you'll have to ask yourself if you prefer a few percent higher performance from Gigabyte's factory OC, or what the ASUS TUF offers.

NVIDIA's alternative options within the GeForce 40 series are also worth considering. Currently priced at $750, the RTX 4070 Ti 12 GB provides slightly lower performance and 12 GB VRAM—I wouldn't opt for it just to save $50, and would rather prefer the 16 GB VRAM peace of mind, even though the difference today is marginal. However, if the Ti non-Super falls to $700 or below, I'd be tempted to lower textures to "high" in two or three games and save $100. On the other hand, you will definitely lose on some resale value, because the used market will certainly appreciate the 16 GB model for longer, which should yield you a better price when you sell it off in a few years.

While RTX 4070 Super is 12 GB, too, its $600 pricing makes it a very attractive option, especially when your focus is 1440p, where the extra FPS of the 4070 Ti Super isn't making that much of a difference. In terms of price/performance you'd be spending +$200 or +33% for a 14% performance improvement. NVIDIA's RTX 4080 ($1200) is too expensive right now, but the RTX 4080 Super promises to change that with a $1000 price point, which could make things interesting, because it might force AMD to lower the price on RX 7900 XTX. If that gets much closer to $900, it could become a viable alternative to 4070 Ti Super, especially for gaming at 4K, where it offers better performance scaling.

Overall, at the end of the day, GeForce RTX 4070 Ti Super is a solid refresh, especially thanks to the VRAM increase, offering increased confidence for the card's longevity in the coming years. RTX 3090 Ti launched for $2000, now you can get identical performance for $800 with the RTX 4070 Ti Super. This new addition from NVIDIA effectively addresses a void within the company's product lineup, specifically targeting the range centered around and slightly above the $800 price point, thereby intensifying competition with AMD's Radeon RX 7900 series offerings.