Nvidia kept the launch of the GeForce RTX 5060 – successor to the popular GeForce RTX 3060 and RTX 4060 series – for Computex. It wasn’t exactly quiet, but the launch was unusual – drivers enabling reviews weren’t released until Monday evening, after sales began. Likely due to backlash over the RTX 5060 still offering only 8GB of memory, four years after the 12GB RTX 3060. As a result, it’s faced a fairly cold reception.
A detailed description of the new technologies brought by Nvidia’s new graphics cards has been compiled by Jan Olšan in the article Blackwell: GeForce RTX 5000 architecture and innovations. It introduces the individual graphics chips Nvidia is using in the new cards, the new GDDR7 memory type, and also covers changes in cache memory, GPU compute units, updates to shaders and tensor cores, as well as improved compute units for ray tracing. The new generation of GeForce also brings support for new standards such as DisplayPort 2.1b and PCI Express 5.0.
With the new generation of cards, DLSS capabilities are evolving as well. It now supports Multi Frame Generation, which allows the generation of up to three intermediate frames, and introduces Nvidia Reflex 2, the second generation of its latency-reduction technology. This includes Frame Warp, which adjusts object and camera motion in rendered frames based on the player’s most recent input device movement data received just before the already rendered frame is displayed. A more detailed explanation can be found in the aforementioned article.
The specifications and suggested prices of the RTX 5060 and RTX 5060 Ti have already been presented in the article GeForce RTX 5060 and 5060 Ti are here: Specs, pricing, availability. After the launch of the RTX 5060 Ti – several models of which we’ve already tested – the more trimmed-down RTX 5060 has now also hit the market, featuring a cut-down version of the chip and 8 GB of memory.
Let me remind you that the older GeForce RTX 5060 Ti is equipped with the smallest chip so far based on the Blackwell architecture, designated GB206. It has a die size of 181 mm² and contains 21 billion transistors. The core contains 4608 stream processors, 144 texture mapping units, and 48 raster units, along with 36 RT cores and 144 tensor cores. The base clock speed of the GPU is 2407 MHz, with a boost of 2572 MHz, and the power limit is set at 180 W. It comes in variants with either 8 or 16 GB of GDDR7 memory, connected via a 128-bit bus. The memory speed is 28 Gb/s, and the total memory bandwidth is 448.0 GB/s.
The newly launched GeForce RTX 5060 retains the number of raster operation units (ROPs), but most of the compute units have been cut down by roughly one-sixth. The chip features 3,840 active stream processors, 120 texture units, and 30 RT cores. Clock speeds are slightly lower – a base clock of 2,280 MHz and a boost clock of 2,497 MHz. The memory runs through a 128-bit bus at 28 Gb/s, resulting in total bandwidth identical to the RTX 5060 Ti. Based on pure specifications, the performance drop for the RTX 5060 compared to the 8GB RTX 5060 Ti should be less than 17%. Against the 16GB variants of the RTX 5060 Ti, the difference in memory-intensive games will, of course, be more significant.
The suggested price for base models of the RTX 5060 is 329 EUR including VAT.
RTX 5060 and 8 GB of memory. Nvidia knows it
The first cards to hit the market were the 16GB variants of the RTX 5060 Ti, followed a few days later by the 8GB version. This week sees the start of sales for the 8GB RTX 5060.
With the 16GB RTX 5060 Ti, there was a full launch with reviews and complete test coverage. But for the 8GB RTX 5060 Ti, Nvidia opted for what is often called a “silent launch” – a quiet release without the usual accompanying reviews. And while manufacturers sent out plenty of review samples for the 16GB version, the same didn’t happen with the 8GB RTX 5060.
This approach is typically used by Nvidia for its lower-end cards, those it doesn’t expect to be well received in reviews. And positive reviews are no longer something 8GB cards can realistically hope for. Lately, reviewers have been going out of their way to prove that games are unplayable on GPUs with 8 GB of memory—by picking the most demanding titles and using settings that clearly exceed what that memory can handle. This intensified especially after it became clear that even the new generation would include RTX 5060 and RTX 5060 Ti models with just 8 GB. The resulting comparisons between the 8GB and 16GB RTX 5060 Ti models focused almost exclusively on this limitation. It was clear the same would happen with the RTX 5060, which is likely why Nvidia chose to downplay the launch of both 8GB models.
There is one difference this time, though—unlike with cards Nvidia wasn’t too proud of in the past, there was no shortage of review samples this time. Previously, reviewers often had to rely on retailers to get a card at all. But for manufacturers, the RTX 5060 series—starting at over €329—is still attractive enough to justify having test samples ready for launch. And so they did offer them for review, just not in advance. Some reviewers grumbled about it, but ultimately accepted that testing would only begin after release.
Then someone in Nvidia’s marketing department came up with the idea to select a few reviewers, reach out to them, and give them drivers ahead of launch—on the condition that they test the cards in a way that would produce graphs matching Nvidia’s own marketing presentations. That meant using selected game titles, specific settings, multi frame generation enabled, and comparisons against older cards that don’t support frame generation—so the RTX 5060’s frame rate gains would look more impressive in charts.
The public reaction wasn’t exactly positive. From a PR and marketing standpoint, I’d say Nvidia did more harm than good. It’s another misstep people are likely to remember for years. The real question is whether anyone outside the hardware enthusiast bubble actually cares—regular customers are likely to buy the RTX 5060 in large numbers anyway. After all, when it comes to price/performance, the RTX 5060 isn’t a bad card—as long as you don’t expect it to do things it can’t, or soon won’t be able to.
Alright, let’s take a look at the first of the RTX 5060 models we’ll be testing.
Asus TUF GeForce RTX 5060 8GB GDDR7 OC Edition
The TUF GeForce RTX 5060 8GB GDDR7 OC Edition, designated as TUF-RTX5060-O8G-GAMING with P/N 90YV0N00-M0NM00, sits at the top of Asus’s RTX 5060 lineup. More affordable alternatives include the triple-fan Prime GeForce RTX 5060 models and the budget-friendly Dual GeForce RTX 5060 – available in both the overclocked O8G variant and the 8G version with reference clocks.
The card comes factory-overclocked. The specified typical boost clock is increased from the reference 2,497 MHz to 2,640 MHz. Actual clock speeds are higher and depend on load and power consumption. You’ll find the clock behavior under load in the sections covering operational characteristics. The GDDR7 memory runs at the standard 28 Gb/s.
The card features two BIOS profiles labeled P mode and Q mode. P mode is the default, and Q mode can be selected via a switch next to the power connector. In the GPU-Z screenshots below, the specifications and data from the default P mode BIOS are shown first, followed by those from Q mode.
GPU-Z version 2.65.1 does not yet recognize the GeForce RTX 5060. Some parameters are missing, and others, like the x16 bus width, are incorrect. Basic parameters such as clock speeds and power limits are identical between both BIOS profiles.
The power limit is increased by only 5 W over the reference 145 W, and can be adjusted in the range of 123–170 W, which corresponds to -21% to +10%. In monitoring software, the card’s three fans are divided between fan 1 and fan 2 interfaces; fan 1 includes the outer fans, and fan 2 is the central fan.
The latest beta version of HWiNFO can extract significantly more data from the card than GPU-Z, you can see the list of currently read sensors in the snapshot from the monitoring.
The latest beta version of HWiNFO reads significantly more data from the card than GPU-Z. The currently available sensors are shown in the monitoring screenshot.
The back of the packaging highlights the features and technologies used by the card. The cooling solution features a heatsink with MaxContact, which should be a more precisely machined base, a backplate with a passthrough cutout, and Axial-tech fans. Durability is supported by dual-ball bearings in the fans, a phase-change thermal pad, a metal frame reinforcing the GPU, and a stainless steel bracket. The card’s compact 2.5-slot cooler and design are intended to fit into the Prime ecosystem. There’s also a note about GPU Tweak III software for monitoring, overclocking, and fan control. The rest includes the usual list of Nvidia technologies found across all GeForce cards.
The accessories include the usual set of leaflets and a quick installation guide. A power adapter is not needed. A nice extra is a double-sided Velcro cable tie, though the support bracket combined with a screwdriver is absent this time.
⠀
