Wednesday, December 25th 2024

NVIDIA GeForce RTX 5070 and RTX 5070 Ti Final Specifications Seemingly Confirmed

Thanks to kopite7kimi, we are able to finalize the leaked specifications of NVIDIA's upcoming GeForce RTX 5070 and RTX 5070 Ti graphics cards.
Starting off with RTX 5070 Ti, it will feature 8,960 CUDA cores and come equipped with 16 GB GDDR7 memory on a 256-bit memory bus, offering 896 GB/s bandwidth. The card is reportedly designed with a total board power (TBP) of 300 W. The Ti variant appears to use the PG147-SKU60 board design with a GB203-300-A1 GPU. The standard RTX 5070 is positioned as a more power-efficient option, with specifications pointing to 6,144 CUDA cores and 12 GB of GDDR7 memory on a 192-bit bus, with 627 GB/s memory bandwidth. This model is expected to operate at a slightly lower 250 W TBP.

Interestingly, the non-Ti RTX 5070 card will be available in two board variants, PG146 and PG147, both utilizing the GB205-300-A1 GPU. While we don't know what the pricing structure looks like, we see that NVIDIA has chosen to make more considerable differentiating factors between its SKUs. The Ti variant not only gets an extra four GB of GDDR7 memory, but it also gets a whopping 45% increase in CUDA core count, going from 6,144 to 8,960 cores. While we wait for the CES to see the initial wave of GeForce RTX 50 series cards, the GeForce RTX 5070 and RTX 5070 Ti are expected to arrive later, possibly after RTX 5080 and RTX 5090 GPUs.
Sources: @kopite7kimi #1, @kopite7kimi #2
Add your own comment

136 Comments on NVIDIA GeForce RTX 5070 and RTX 5070 Ti Final Specifications Seemingly Confirmed

#126
nguyen
KritComparing GTX 1080Ti to RTX 4090 is complete mismatch and dishonor. GTX 1080 Ti has way, way better price/performance ratio than RTX 4090 in straight comparision. And 8800 GTX / 8800 GT was, way better than even GTX 1080Ti / GTX 1070...... Performance gains were tremendous for almost the same price! Where's the RTX 4090 is complete piece of shit in this region.
4090 holds the record for the longest uncontested GPU in the entire gaming GPU history, but sure let just go with your made up metrics for why it's bad :roll:.

BTW I owned 8800GT and 1080 Ti too, no particular fondness of them because I just buy the next best GPU anyways
Posted on Reply
#127
Prima.Vera
nguyen4090 holds the record for the longest uncontested GPU in the entire gaming GPU history, but sure let just go with your made up metrics for why it's bad :roll:.

BTW I owned 8800GT and 1080 Ti too, no particular fondness of them because I just buy the next best GPU anyways
That's because back in the day they used to release video cards every year, not every 3 years like now. :)
Posted on Reply
#128
nguyen
Prima.VeraThat's because back in the day they used to release video cards every year, not every 3 years like now. :)
Meaning back then enthusiasts will buy new GPU every year (I did anyways), so the prices stacked up :cool:
Posted on Reply
#129
Onasi
@Krit
Bruh, the context is kinda important here. Me and @nguyen were bantering about chips, SKUs and potential stack performance and deltas between both current and future gen cards. We both acknowledged high pricing for incoming cards in that discussion too. Not one of us brought up price/performance as a metric. You invented an argument to go against for no reason. Chill out.
Posted on Reply
#130
wheresmycar
Intel introduced the 192-bit 12GB B580, priced at $250, setting a solid foundation for the "lower" performance tier.

The minimum expectation for Nvidia and AMDs Next Gen "mid" perf tier GPUs is nothing short of 256-bit memory interface and 16GB of VRAM. Anything less, would once again, demonstrate a clear disregard for consumer expectations and the need to provide adequate headroom for advancing graphics performance.

Similar to how the standard 4070 was criticized as essentially a masked 4060 T/S, the 5070 appears to be little more than a disguised 5060 T/S.

Putting the 5070 and 5070 TI to the side for the mo, I was hoping to grab a possible 5070 TI SUPER somewhere in the $800 region to top my RTX 3080............ STOOOOPID WISHFUL THINKING!!
Posted on Reply
#131
gffermari
You shouldn't stick to spec numbers. 256 or 192bit doesn't say anything to anyone.
All it's about performance (per cost).
Posted on Reply
#132
Dahita
DaworaPpls only see one GPU, its like there is no other otions..

How about 5060Ti 16GB or 5070Ti 16GB?
5060Ti for 1440p? That would be their low end GPU, suitable mid-term for 1080p.

The 5070TI would be a good option. What are the chances it will be released at $500? None :( However, you make me realize that I bought my 2070 SUPER brand new $584.99 at Microcenter in 2019. If I adjust the price with inflation, it represents $724.26 today. So my request is not valid after all. Guess I'm going to have to fork out the money!
Posted on Reply
#133
Dawora
TumbleGeorge-30% too small number of cores to have better ratio.
U are totaly wrong..

5070Ti will be very close to 4090
Posted on Reply
#134
xnery.007
I plan to purchase the 5070 Ti, as it seems like a very interesting card considering it uses the GB203 (yes, even though it's cut down compared to the 5080) and has 16GB of VRAM. DLSS4 and other features will be a nice bonus.

I estimate the price at $849, which would be a $50 increase compared to the 4070 Ti SUPER. During the launch of the 4xxx SUPER variants, Nvidia acted very reasonably, as it was their last launch in that series. Based on that, I believe Nvidia will follow a similar approach this time.

Additional reasons:
  • It still uses a 256-bit bus
  • The 4N(P) node is more advanced than 5nm, but it offers a good price per wafer and excellent yield
  • Inflation has stabilized
  • GDDR7 isn’t significantly more expensive than GDDR6X
Posted on Reply
#135
Legacy-ZA
DaworaU are totaly wrong..

5070Ti will be very close to 4090
Nah, the leather jacket man is not that kind, it will most likely land between a 4080S and 4090. I am disappointed that it doesn't have 24GB of VRAM though, 16GB being entry for high resolution gaming, that could have been done with 3GB modules, easily.

At this point, the VRAM issue, is far more than just a nuisance, it's deliberate sabotage out the gate, making sure the GPU's don't last for more than one generation.
Posted on Reply
#136
gffermari
The 5070Ti will be +10-15% faster than a 4080. The 16GB is fine for the card.
My issue is the 5080 is heavily butchered. Even more than the 4080.
It's a joke that the 5090 gets 32GB and the second best half of it.
Most likely the greens will release a 5080Ti with 24GB but it appears the x80s are not a high end performance segment anymore (the 3080 is excluded).
Posted on Reply
Add your own comment
Dec 27th, 2024 08:06 EST change timezone

New Forum Posts

Popular Reviews

Controversial News Posts