NVIDIA GeForce RTX 5060 Ti Full Specs Leaked — What We Know Ahead Of The Launch

Just two days ahead of its official launch, the full specifications of NVIDIA’s GeForce RTX 5060 Ti have surfaced online, courtesy of a leak through GPU-Z. The leak provides a detailed view of the card’s architecture and performance attributes, putting an end to speculation surrounding the mid-range RTX 50-series model.

The particular version, reportedly the standard MSRP variant, comes equipped with 16GB of GDDR7 memory. The memory runs at a clock speed of 1750 MHz, which translates into an effective speed of 28 Gbps. Thanks to a 128-bit memory bus, the card can achieve a total memory bandwidth of 448 GB/s. Samsung is confirmed as the sole memory supplier for this model, with no signs of Hynix chips being included for now.

The card features 4608 CUDA cores, alongside 144 texture mapping units (TMUs) and 48 render output units (ROPs). The GPU base clock is rated at 2407 MHz, with a boost clock reaching 2572 MHz. In terms of connectivity, the card uses the PCIe 5.0 interface, although it operates over just eight lanes (x8), which may have implications in systems with heavy GPU-to-CPU traffic.

One important detail that has emerged from the leak is the card’s fixed power configuration. The RTX 5060 Ti carries a locked thermal design power (TDP) of 180 watts, with no option to manually adjust it. The BIOS for this particular model was finalised on 16 March, suggesting it is production-ready and close to distribution.

The graphics card is currently running on a pre-release driver, version 575.94. This is expected to be the base driver for initial reviews, while the official WHQL-certified release is likely to follow at launch.

NVIDIA will lift the review embargo on Apr. 15. Early reviews will focus exclusively on the 16GB variant, while the 8GB model, though in the pipeline, is not expected to feature in initial benchmarks.

It is notable that there is no Founders Edition version of the RTX 5060 Ti planned.