With the launch of GeForce RTX 5070 done and dusted, attentions are now firmly turning to Nvidia’s budget-class graphics cards. The company hasn’t formally announced RTX 5060 Ti, 5060, or 5050 but this hasn’t stopped rumourmongers from sharing alleged specifications for the three SKUs.
In a series of X posts, infamous Nvidia leaker kopite7kimi shared details on GeForce RTX 5060 Ti, 5060, and RTX 5050. I’ve collated their claims into a specs table below.
| GeForce | RTX 5060 Ti | RTX 5060 | RTX 5050 |
|---|---|---|---|
| GPU | GB206 | GB206 | GB207 |
| CUDA cores | 4,608 | 3,840 | 2,560 |
| Memory | 8/16GB | 8GB | 8GB |
| Mem. type | GDDR7 | GDDR7 | GDDR6 |
| Mem. interface | 128-bit | 128-bit | 128-bit |
| Board power | 180W | 150W | 130W |
If kopite7kimi proves reliable then GeForce RTX 5060 Ti and 5060 will mark the debut of Nvidia’s GB206 GPU. Though they’ll share he same chip, the latter graphics card will feature a cutdown die á la RTX 5070 Ti. Note that RTX 5060 has 17% fewer CUDA cores at its disposal, as this same shortfall will apply to RT and Tensor cores too.
Both GeForce RTX 5060 Ti and 5060 pack GDDR7 VRAM, though kopite7kimi doesn’t confirm the speed of their memory modules. If these cards mirror their siblings, then they could run at 28Gb/s netting a total bandwidth of 448GB/s. This would represent a sizeable increase over their forebears, as RTX 4060 Ti and RTX 4060 top out at 288GB/s and 272GB/s, respectively.
kopite7kimi echoes prior whispers that GeForce RTX 5060 Ti will be available with up to 16GB of VRAM. However, RTX 5060 won’t receive the same treatment and have to make do with just an 8GB buffer. It’s disappointing to see budget gamers saddled with constrained capacity again, particularly after RTX 4060 received plenty of criticism for this upon release and thereafter.
Like other GeForce RTX 50 Series graphics cards, both GeForce RTX 5060 Ti and RTX 5060 will require more power than their predecessors. The generational shift isn’t massive at 180W (+15W/20W) and 150W (+35W) but I’m curious to see how much more performance those extra watts will bring. While we’re on the subject of power, some suggest that Nvidia won’t mandate 16-pin connectors on these SKUs.
GeForce RTX 5050 is an altogether different beast from its 60 class siblings. It will apparently rock a GB207 GPU, the only model in the RTX 50 Series to do so. As the least-powerful chip in the stack, it’ll only pack 2,560 CUDA cores and a relatively small amount of RT and Tensor cores too. It also carries the smallest power footprint of its family, at 130W.
Curiously, Nvidia is opting for GDDR6 VRAM rather than GDDR7 in the case of RTX 5050, likely in a bid to reduce costs. There’s nothing inherently wrong with this strategy, as the memory standard serves Radeon RX 9070 XT an RX 9070 just fine. However, it’ll only have 8GB of the stuff to play with, hampering its ability to handle higher-resolution textures, ray traced effects, and more.
Given how lukewarm the reception of GeForce RTX 5070 is, these specifications don’t suggest a turning point for Nvidia’s public perception. Nothing’s for certain until we can test the cards’ mettle, of course, and you should take all claims above with a pinch of salt until Team Green reveals what it has in the works.
Regardless, I’m excited to see how GeForce RTX 5060 and Radeon RX 9060 Series cards play off against one another. AMD could play another blinder, securing the market share it so desperately needs. However, there’s still a chance that Nvidia could surprise us all with genuinely high-value options.
The battle for the budget market should commence imminently, as an RTX 5060 Ti reveal event could be just over the horizon at GTC. Make sure you’re following Club386 on Google News to say in the loop regarding any developments as they happen.
