To date, all of Nvidia’s of GeForce RTX 40-series graphics cards based on the Ada Lovelace architecture target enthusiasts with rather deep pockets. But the lineup is expanding to the midrange with products that should carry lower price tags. According to MyDrivers, Nvidia’s GeForce RTX 4060 Ti will be one of those midrange boards that will retail for sub-$500. Meanwhile, it will offer performance akin to that of Nvidia’s GeForce RTX 3070, says well-known hardware leaker @kopite7kimi. 

The cheapest GeForce RTX 40-series graphics card introduced to date is the GeForce RTX 4070 Ti that features a manufacturer suggested retail price (MSRP) of $799, with performance that’s about equal to the previous generation RTX 3090 (non-Ti). Nvidia yet has to disclose prices for its GeForce RTX 4070 (non-Ti) and GeForce RTX 4060 Ti parts, but since these parts will sit below the RTX 4070 Ti, the model RTX 4060 Ti will naturally retail at a lower price. Nvidia will make final decisions about the MSRPs of its products closer to launch, but general feedback on the 4070 Ti and 4080 pricing has been quite harsh.

As far as performance goes, Kopite7kimi suggests the 4060 Ti will only match the 3070, which would be pretty poor, frankly. With the 4070 Ti matching the 3090, the vanilla 4070 ought to at least match the 3080, if not the 3080 12GB / 3080 Ti. That would put the 4060 Ti at either 3070 Ti levels or even on par with the 3080 10GB.

Nvidia’s RTX 3060 Ti currently ranks as one of the best graphics cards for its price bracket, but it’s due for replacement. The 4060 Ti will use a newer microarchitecture, and performance is likely to improve as well. But Nvidia has pushed prices up on every GPU class for the 40-series so far, and it wouldn’t be surprising — though it would be disappointing — to see the future 4060 Ti priced $100 higher than the previous generation card.

Specification-wise, the GeForce RTX 4060 Ti is a rather curious product. Based on unofficial information, the GeForce RTX 4060 Ti will rely on the AD106 GPU with 4352 CUDA cores mated with 8GB of 18 Gbps GDDR6 memory using a 128-bit interface. The new part would feature considerably fewer stream processors than the alleged GeForce RTX 4070 with 5888 CUDA cores, plus 33% less memory, which suggests a substantial performance gap between the two. About the only bright spot is that the RTX 4060 Ti is projected to consume 160W, which would be very low by today’s standards. 

Furthermore, the GeForce RTX 4060 Ti is expected to feature a 43% lower memory bandwidth compared to that of the GeForce RTX 4070, as well as a 36% lower DRAM bandwidth than its direct predecessor, the GeForce RTX 3060 Ti with GDDR6 memory. The larger L2 cache should help, relative to the 30-series, but it would still be a wide gulf between the 4060 Ti and the 4070. Coming with just 8GB, the same as the 3060 Ti, 3070, and 3070 Ti (but less than the 3060) is also problematic. How this will affect its performance in cases of high resolutions and/or with antialiasing enabled is something that remains to be seen.

Nvidia will of course tout the benefits of DLSS 3 and Frame Generation yet again, but the further down the stack we go, the less beneficial FG becomes. Our testing to date indicates the best experience comes when the base FPS is already in the 60+ fps range. That’s easy enough for cards like the RTX 4090 and 4080 to hit at 4K, but an 8GB GPU may struggle to reach that level of performance even at 1440p without dropping some of the other settings.

In any case, since we are talking about leaked details from unofficial sources, take the information with discretion since it may not be completely accurate. We’re still looking forward to testing the future 4070 and 4060 Ti to see how they stack up, and hopefully prices will start to approach the true midrange sector.