Of course, no 12GB 5060. I guess that rumour was fake. I'm never getting a new PC at this rate
DegustatoR
Legend
We don't know yet, there are rumors.
Isn't videocardz trustworthy?
DegustatoR
Legend
This one is about final 5060 Ti spec, 5060 there is just what's known at the moment.
It is worth remembering that IHVs plans always change up until they give you the official announcement and specs.
Same Videocardz have these from recently:
Then there is still a chance
It seems like it. They are being used on the RTX5090M and the "RTX PRO" workstation cards.
DegustatoR
Legend
NVIDIA Reportedly Narrows Down GeForce RTX 5060 Ti Launch to April 16
Last week, we heard rumors about NVIDIA delaying its launches of GeForce RTX 5060 Ti and RTX 5060 by a couple of weeks. Initially, PC hardware watchdogs anticipated a product unveiling before GTC 2025 kick-off time. Industry insiders did not fully disclose the reasons behind Team Green's revised...
www.techpowerup.com
arandomguy
Veteran
Samsung had previously announced 3GB modules for 2025Q1.
From what I can tell official catalogue listings for Hynix, Micron and Samsung only still list 16Gb/2GB.
So it's going to depend on what you mean by available, see below.
Both have been announced but neither is available for purchase yet. So it is unclear what the shipping and distribution status of parts using 3GB chips is at this stage.
Last edited:
arandomguy
Veteran
NVIDIA Reportedly Narrows Down GeForce RTX 5060 Ti Launch to April 16
Last week, we heard rumors about NVIDIA delaying its launches of GeForce RTX 5060 Ti and RTX 5060 by a couple of weeks. Initially, PC hardware watchdogs anticipated a product unveiling before GTC 2025 kick-off time. Industry insiders did not fully disclose the reasons behind Team Green's revised...www.techpowerup.com
The videocardz article suggests that launch day reviews on the 16th will only be 16GB models. Which let's just say at best should raise some eyebrows if that turns out to be true...
The new software feature launch, G-Assist, today requiring 12GB or higher is not going to help with the optics here.
Lol 5060 not even enough memory to run the AI that tells you to turn textures down
But RTX 3060 12 GB is happy to run it while RTX 3080* owners can'tLol 5060 not even enough memory to run the AI that tells you to turn textures down
*original, not the later released 12 GB version
Yea I'd rather have a 3060 that a 4060 or 5060. Those cards don't have enough memory to make good use of new features like framegen anyway. Strange times.But RTX 3060 12 GB is happy to run it while RTX 3080* owners can't
*original, not the later released 12 GB version
DegustatoR
Legend
Why would it raise any eyebrows considering that these SKUs will have different MSRPs?
Also if they do want to pivot 5060 into launching with 12GB it makes sense to prioritize the 16GB 5060Ti for reviews.
In both cases the 8GB versions would be cheaper but with a very obvious drawback.
G-Assist has some interesting features, like ability to analyze game performance and then suggest changes based on it. The problem is, it's heavy while doing so, lowering game performance and then basing it's suggestions on that lowered performance. 
As reported by TechBBS user, he has set up his Indiana Jones and the Great Circle for speedruns and it runs near 200 FPS (Core i9-14900K + RTX 4090). When he tried using G-Assist to see what it says, the FPS dropped to 55-60 and report on suggested changes to settings was based on that performance, not the near 200 it actually runs.
As reported by TechBBS user, he has set up his Indiana Jones and the Great Circle for speedruns and it runs near 200 FPS (Core i9-14900K + RTX 4090). When he tried using G-Assist to see what it says, the FPS dropped to 55-60 and report on suggested changes to settings was based on that performance, not the near 200 it actually runs.
Last edited:
arandomguy
Veteran
It should be a cause of concern if they are only sampling 16GB for reviews but selling the 8GB variant with the same launch date. Not to mention the misleading marketing if MSRP messaging is for the 8GB variant but all launch review data is for the 16GB variant. Again we are assuming that the videocardz article is true.
This would be no different than launching a GPU with $X MSRP but then only sampling OC variants with +10% performance that cost more for reviews. It would actually be worse since unlike clock speeds there is no way to "simulate" an 8GB card vs 16GB.