You might at first be confused as to why NVIDIA is calling this Geforce GT 710 a “new” card. After all, it’s been on the market for more than two years now using two different architectures, namely Fermi (GF119) and Kepler (GF208). This “new” GT 710 is a higher-clocked rebrand of the GF208 model, equipped with 192 CUDA cores, 16 TMUs, 8 ROPs and between 1-2GB of DDR3 memory running at 900MHz. While this isn’t an exciting thing in the least, it is a sign that we’re probably moving to a new lineup of NVIDIA GPUs this year that finally bring the Maxwell architecture to the low-end parts. I’m not sure if that’s still going to be the Geforce 900 series though, because the current lineup is horribly disjointed, with the Geforce 700 series only meeting up halfway with the GTX 750 Ti.
Having Maxwell at the low-end is necessary because those cards will enable HTPC owners to upgrade to something capable of smoothly playing HEVC H.265 content using the GPU, rather than trying to make their poor CPU do all the hard work. Better performance is another benefit, and a Maxwell-based version of this GPU could be interesting if equipped with GDDR5 memory – not for gaming, mind you, but for running CUDA plugins and doing some GPGPU work. As office workhorses, these won’t do too badly.
The slightly faster GT 710 is available today in various designs from Zotac, EVGA, and ASUS, which share a starting price of about U.S. $40 (approx. R650*). You can identify them because they have a core clock of 954MHz, a 54MHz overclock compared to the standard GPU. It may even boast better performance than the GT 720, which has the same hardware setup with a lower core clock of 797MHz.