The GeForce RTX 5060 will be a mid-range graphics card based on the NVIDIA Blackwell architecture. We know it will be based on Graphics core GB207and that it will arrive after the launch of GeForce RTX 5090, GeForce RTX 5080 and GeForce RTX 5070. This means that the launch is still a few months away.

NVIDIA has not yet officially confirmed the GeForce RTX 5060, but in recent months I have seen numerous information and leaks related to this graphics card and have done some work collecting and curating content for the integration. most reliable and interesting details in this article, which is a necessary update to the one I posted last April.

All the information you will see below comes from reliable sources and makes a lot of sensebecause as I said before, I carefully filtered it to leave out nonsense and information with little or no basis. But since it is not officially confirmed, this could change, although the truth is that it would be unlikely at this point.

Possible specifications of GeForce RTX 5060

• Graphics core GB207 made on TSMC 4nm node.
• 32 SM units.
• 4,096 shaders at speeds between 2.8 GHz and 3.2 GHz.
• 128 texturing units.
• 48 raster units.
• 32 cores of the fourth generation RT.
• 128 fifth generation tensor cores.
• 128-bit bus.
• 8 GB or 12 GB of GDDR7 memory at 28 Gb/s (448 GB/s bandwidth).
• 32 MB L2 cache.

If these specs are confirmed, the GeForce RTX 5060 will have a significant increase in shaders compared to the GeForce RTX 4060, as it has 3,072 shaders. The difference in favor of the former would be 1,024 shadersand obviously it will also have more texture units, more tensor cores and RT cores, and more L2 cache memory.

Having a larger number of shaders running at a higher frequency means more brute force, and with more RT cores and more tensor cores, the GeForce RTX 5060 will perform better than the GeForce RTX 4060 when working with ray tracing and AI. We must not forget that these cores will also be more advanced and will introduce improvements derived from the Blackwell architecture.

The memory bus will remain at 128 bits, but the GeForce RTX 5060 will use GDDR7 memory, allowing it to significantly increase bandwidth. This will achieve 448 GB/swhile the GeForce RTX 4060 only reaches 272 GB/s. Memory capacity could stay at 8GB or increase to 12GB using Samsung’s new 3GB chips.

The L2 cache will also increase in the GeForce RTX 5060, which is an important fact because it has a larger one the number of accesses when searching this cache increases, and thereby reducing the dependency on the graphics memory subsystem. The 32 MB of L2 would be enough to achieve optimal results at 1080p and at 1440p it lasts much better than the 24 MB GeForce RTX 4060.

If the GeForce RTX 5060 arrives with 8GB of graphics memory, its target resolution would be exactly that, 1080p. A model with 12 GB of graphics memory would be ready to work optimally with a resolution of 1440p. Note that this doesn’t mean the 8GB version won’t be able to move games at 1440p, I’m just saying that amount of graphics memory It is not an ideal level for the said resolution.

How the GeForce RTX 5060 will use the Blackwell architecture its shader performance at the same frequency should be higher on the GeForce RTX 4060. NVIDIA will also improve the performance of tensor cores and RT cores and may introduce technologies exclusive to this new architecture. We must also not forget the difference that the jump to TSMC’s 4nm node will make, which reduces transistor size and improves efficiency.

Possible performances and equivalences

A GeForce RTX 5060 configured to the specs I listed above would have the approximate performance 26.21 TFLOPs in FP32. This number far exceeds the GeForce RTX 4060’s 15.11 TFLOPs and puts it above the GeForce RTX 4060 Ti, but is below the GeForce RTX 4070’s 29.15 TFLOPs.

TFLOPs in FP32 allow us to compare the performance of two graphics cards, but may not represent actual gaming performance. This GeForce RTX 5060 would almost certainly be more powerful than GeForce RTX 4070 in games, but it would be below GeForce RTX 4070 SUPER, they always talk about 1080p resolution.

Ray tracing can be different. Thanks to the new fourth generation of RT cores, the GeForce RTX 5060 It could be closer to GeForce RTX 4070 SUPER. This will largely depend on the amount of graphics memory NVIDIA chooses to integrate, as ray tracing consumes a lot of memory and not enough can negatively impact performance.

For power and L2 cache GeForce RTX 5060 would be able to move games at 1440p resolution no problem, but it will depend on whether it comes to the market configured with 8 GB or 12 GB of graphics memory. In the first case, the ideal resolution would be 1080p and in the second 1440p. There are many games that run optimally in 1440p with 8 GB of graphics memory, but when we move to this resolution, there is a clear trend towards 12 GB.

NVIDIA has the final say, but the truth is that with Samsung’s 3GB GDDR7 memory chips it is very easy for GeForce RTX 5060 configuration with 12 GB VRAM and 128-bit bus.

Release date and possible price

As I said at the beginning of the article, the launch of GeForce RTX 5060 will happen after the arrival of GeForce RTX 5070, which in turn will be launched after GeForce RTX 5080 and GeForce RTX 5090. This data set brings us to first half of 2025and the truth is, it’s hard to bet on a specific month right now.

With all the information we have right now, I think it’s most likely that the GeForce RTX 5060 will arrive between March and June. It will all depend on what NVIDIA does with the GeForce RTX 5070 and the rush that will be depending on the stock of the GeForce RTX 4060.

In terms of price, the GeForce RTX 5060 should be in between 249 and 299 dollarsthe numbers they are likely to become at the exchange rate in Spain and after taxes are applied 299-349 euros, approximately.

What power supply and processor do I need to power a GeForce RTX 5060

This graphics card will be more powerful than the GeForce RTX 4060 and will also have a higher consumption, as it could range between 150 and 170 watts, depending on the final configuration that NVIDIA uses. The difference would be 35 to 55 watts more on the GeForce RTX 5060 quite reasonable data that will not force us to invest in a very powerful resource.

If we have a typical computer consisting of Ryzen 5 7600, 32 GB DDR5, PCIe NVMe SSD, six fans and 240mm AIO liquid cooling, this will be enough for us. 500 watt power supply to move the GeForce RTX 5060, if the mentioned source will be of minimum quality and have a 16-pin power connector, which will be necessary for the reference model.

This configuration should a average consumption 350 watts under load if the GeForce RTX 5060 has a TGP of 170 watts. With the price difference between 450 watt and 500 watt sources, it is not worth investing in a lower one, so I recommend the 500 watt one.

It goes without saying that if we have a PC with a processor whose consumption is higher, we will need a more powerful resource. For example, with an Intel Core i9-13900K, consumption would increase 523 watts under load, and in this case it would be advisable to fit a 600 watt power supply to have some margin.

As for the recommended processor to power the GeForce RTX 5060, the recommended minimum would be Ryzen 5 5600 or Intel Core i5-12400Fand Ryzen 5 7600 or Intel Core i5-12600KF, or Ryzen 7 5700X3D would be ideal. If you have questions about what processor you need for each graphics card and want to delve deeper into this topic, don’t miss this tutorial.

AI generated images.