Nvidia's RTX 5060 Ti is the latest Blackwell GPU to hit the market – and yes, there is still an 8GB model launching in 2025… This review is focused on the 16GB SKU though, hitting the market with a £399/$429 MSRP, a price reduction compared to its predecessor. With 4608 CUDA cores, a 128-bit memory interface, and new GDDR7 memory, the 5060 Ti 16GB targets the mainstream gamer, aiming to deliver strong 1080p and playable 1440p performance. Today we find out exactly what it is capable of.
With no Founders Edition for either RTX 5060 Ti model, Nvidia is relying on its partners for this launch. We were sent two cards, both of which are featured in today's review. The majority of my game testing was conducted using the Palit RTX 5060 Ti 16GB Infinity 3, an MSRP card that runs at reference clock speeds. I also have the Gigabyte Aorus Elite, a much more premium model, which we also tested today.
Timestamps
00:00 Intro
00:49 Meet the 5060 Ti
01:43 Test setup
02:49 Alan Wake 2
03:16 Black Myth: Wukong
03:53 Cyberpunk 2077
04:23 Dragon Age: The Veilguard
04:44 Final Fantasy XVI
05:06 Ghost of Tsushima
05:33 Horizon Forbidden West
06:15 Kingdom Come Deliverance 2
06:37 Plague Tale: Requiem
06:55 Ratchet and Clank
07:15 Senua’s Saga: Hellblade 2
07:40 Starfield
08:00 Performance summary – 12 game average
09:00 Cost per frame analysis
10:11 Ray tracing benchmarks
12:46 DLSS 4, Multi Frame Gen testing
15:43 Meet the Palit Infinity 3
16:12 Gigabyte Aorus Elite
16:55 Clock speed
17:27 Thermals and acoustics
18:33 Power draw + efficiency
20:15 Closing thoughts

|
RTX 5080 |
RTX 5070 Ti |
RTX 5070 |
RTX 5060 Ti |
RTX 4060 Ti |
Process |
TSMC N4 |
TSMC N4 |
TSMC N4 |
TSMC N4 |
TSMC N4 |
SMs |
84 |
70 |
48 |
36 |
34 |
CUDA Cores |
10752 |
8960 |
6144 |
4608 |
4352 |
Tensor Cores |
336 |
280 |
192 |
144 |
136 |
RT Cores |
84 |
70 |
48 |
36 |
34 |
Texture Units |
336 |
280 |
192 |
144 |
136 |
ROPs |
112 |
96 |
80 |
48 |
48 |
GPU Boost Clock |
2617 MHz |
2452 MHz |
2512 MHz |
2572 MHz |
2535 MHz |
Memory Data Rate |
30 Gbps |
28 Gbps |
28 Gbps |
28 Gbps |
18 Gbps |
L2 Cache |
65536 KB |
49152 KB |
49152 KB |
32768 KB |
32768 KB |
Total Video Memory |
16GB GDDR7 |
16GB GDDR7 |
12GB GDDR7 |
8/16GB GDDR7 |
8/16GB GDDR6 |
Memory Interface |
256-bit |
256-bit |
192-bit |
128-bit |
128-bit |
Memory Bandwidth |
960 GB/Sec |
896 GB/Sec |
672 GB/Sec |
448 GB/Sec |
288 GB/Sec |
TGP |
360W |
300W |
250W |
180W |
160W |

First, a quick spec recap. RTX 5060 Ti marks the introduction of a new Blackwell die – GB206, and this is a full implementation. Comprised of three Graphics Processing Clusters (GPCs), each holds up to eight Texture Processing Clusters (TPCs), with a total of 18. Each TPC is home to two Streaming Multiprocessors (SMs), giving us 36, and each SM still holds 128 CUDA Cores, meaning the RTX 5060 Ti has a total of 4608 shaders. We also find 36 RT cores, 144 Tensor cores, 144 Texture Units, and 48 ROPs.
This time around, however, there's no node-shrink, and GB206 remains fabricated on TSMC's N4 node, as per the RTX 40 series. As such, rated clock speed has not stepped forwards this generation, with the RTX 5060 Ti rated at 2572MHz, within 40MHz of the RTX 4060 Ti.
The memory configuration is similar to its predecessor, too. We still find 8 or 16GB of memory operating over a relatively narrow 128-bit interface, the primary difference being the jump to 28Gbps GDDR7 modules, increasing memory bandwidth to 448 GB/s. L2 cache still comes in at 32MB.
Lastly, the RTX 5060 Ti features a 180W TGP. This is something we focus on closely in this review, using our enhanced GPU power testing methodology, so read on for our most detailed power and efficiency testing yet.
The post
Nvidia RTX 5060 Ti 16GB Review ft. Gigabyte & Palit first appeared on
KitGuru.