The RTX 4060 Ti graphics card will have two variants, offering 8 GB (like the non-Ti version) or 16 GB of VRAM. AMD's mid-range model, also featuring 8GB of memory, is expected to be released by the end of May.

Team Green

According to official data, the RTX 4060 Ti series will provide up to 22 TFLOPS of FP32 single-precision computing performance, while the shader performance of the non-Ti RTX 4060 is expected to be around 15 TFLOPS. NVIDIA showed this data compared to the RTX 3060 Ti and RTX 3060 series which offer 16 TFLOPS and 13 TFLOPS respectively.

The GeForce RTX 4060 Ti model will have a TDP of 160W for both variants, while the average power consumption should be around 140W during gaming. The RTX 4060 will launch with a TDP of 115W, while typical power consumption during a gaming session should be around 110W.

  RTX 4060 Ti RTX 4060
CUDA cores 4352 3072
RT cores 3rd gen, 51 TFLOPs 3rd gen, 35 TFLOPs
Tensor cores 4th gen, 353 TFLOPs 4th gen, 242 TFLOPs
Memory bandwidth 288 GB/s 272 GB/s
L2 cache 32 MB 24 MB
Average gaming power 140 W 110 W
Video playback power 13 W 11 W
Idle power 7 W 7 W
TGP 160 W 115 W

As for specs, the RTX 4060 Ti equips 18 Gbps 128-bit GDDR6 memory, while the non-Ti variant uses 17 Gbps chips. This means that the non-Ti variant will have a lower maximum bandwidth of 272 GB/s compared to the 288 GB/s of the Ti series.

NVIDIA also showed performance cards, which show a noticeable improvement with DLSS3 (Frame Generation). However, it is always advisable to wait for third-party tests to gauge the actual performance improvement, as has been the case in the past.

NVIDIA has confirmed that the GeForce RTX 4060 will cost $299, making it the cheapest desktop card of the latest generation of GPUs. Both the RTX 4060 and RTX 4060 Ti with 16 GB of VRAM will launch in July, while the 8 GB Ti model will launch on May 24th.

Hardware Unboxed conducted tests on the RTX 4060 Ti and revealed a modest improvement (approximately 5% to 10%) in performance when compared to its forerunner, the RTX 3060 Ti.

Team Red

AMD's Radeon RX 7600 will be based on the Navi 32 XL graphics processor, with 32 compute units or 2048 stream processors paired with 128 texture units and 32 raster units that can reach a default gaming clock frequency of 2250 MHz up to 2655 MHz. The GPU will have a 128-bit GDDR6 memory interface that offers a bandwidth of 288 GB/s, via a PCIe 4.0 x8 interface, with 8 GB of memory.

GPU Navi 33
Stream processors 1792
RT cores 28
Base clock 2250 MHz
Boost clock 2655 MHz
Memory  8 GB GDDR6
Memory bus 128-bit
Memory clock 18 GHz effective
Memory bandwidth 288 GB/s
TDP / TGP 175 W
Interface 8x PCIe 4

On May 25th, AMD is set to launch the non-XT variant of its Radeon RX 7600 GPU, with a starting price of $270. Additionally, it is worth noting that there was a last-minute price reduction from the originally announced price of $300.

This price cut positions the RX 7600 as the current reigning champion in terms of performance per dollar, as demonstrated by Hardware Unboxed in its review.