New leak reveals alleged specs for GeForce RTX 2050, entry-level GTX 1150 cards

Ahead of CES 2019 in Las Vegas, the graphics card rumors just keep on coming. In news that will make fans of Nvidia’s gaming processors giddy, some alleged specifications for the GeForce RTX 2050 graphics card and the rumored entry-level GTX 1150 card have been revealed by a known leaker going by the name TUM_APISAK on Twitter.

NVIDIA Graphics Device 14 CU (896 cuda?)
2050/1150 ? ????
OpenCL score
1050ti – 84000
1050 – 73000 pic.twitter.com/S8a6NdWlDW

— APISAK (@TUM_APISAK) December 24, 2018

As usual, these latest leaks don’t officially reveal much from Nvidia, but again confirm that the RTX 2050 is real and could be coming soon. The listed Geekbench specifications hint at a total of 30 compute units, a maximum processing frequency of 1.2 GHz, and a total of 6GB graphics memory onboard the RTX 2050. As for the OpenCL score, it comes in at 211,896, which looks to be fairly high enough for powerful gaming machines.

The same leak also reveals some more information on what could be Nvidia’s GTX 1150 entry-level processor. Rumored to be the replacement to the GeForce GTX 1050 Ti, the Geekbench specifications simply list the chipset as “Graphics Device.” Given previous talks of all new non-RTX GPUs being pushed to the GeForce 11 series, it is likely built on Nvidia’s Turning architecture and comes in with specifications close to this year’s GeForce GTX 1050 Ti. The OpenCL score also comes in at 114,206, and the compute units at 14, with a maximum frequency of 1.56 GHz.

Though other details like the memory standard are not available from this listing, the overall OpenCL benchmark on the GTX 1150 hints at a bump in performance. Compared to the 1050Ti and 1050, the scoring goes up from 84,000 and 73,000 to 114,206. That would be a jump of 30,000, and 41,000 between generations.

Ahead of CES, these leaks and rumors are shedding more light on Nvidia’s plans for 2019. Earlier on Wednesday, December 26, filings from the Eurasian Economic Commission showed up to 40 different variations of the GeForce RTX 2060. Those graphics cards could likely come in three different variants with each having alternate levels of video memory, either 6GB, 4GB, and 3GB. The SKUs will also support the GDDR6, GDDR5X, and GDDR5 video memory architecture, which is a departure from GDDR6 memory.

It is currently not clear how the GTX 1050 successor or the RTX 2050 will be sold or branded, or at what pricing. You can expect to hear more about these graphics cards during Nvidia’s CES 2019 conference, scheduled for January 7 at 8 p.m. PT.