Following the excitement of CES 2025, NVIDIA shared additional details about the new 50-series GPUs and what their Blackwell-based architecture can deliver.

We do a lot of benchmarking, and Nvidia is bringing some new features to its RTX 50-series GPUs that require changes in how to collect performance data. Here's what you need to know about Nvidia's "best practices" as well as our plans for future testing of GPUs.

On January 6, 2025, NVIDIA Founder and CEO Jensen Huang took to the CES 2025 stage to unveil the RTX Blackwell, a new series of graphics cards from the company. This family of GPUs is the RTX 50-series of cards:

For many, specs will be the main point of comparison between the RTX 5090 and RTX 4090. That makes sense given that the Blackwell GPU will use a shiny new architecture paired with 32GB of faster GDDR7 VRAM and enhanced AI abilities, but it's worth noting that on-paper figures don't always translate performance-wise.

Nvidia provided in-depth details at its recent Editors Day on the Blackwell architecture, specifically the consumer variants that will be used in the RTX 50-series GPUs. Here's the technical overview of what's going on under the hood with the upcoming graphics cards.

The GeForce RTX 50 Series, a new generation of GPUs powered by the NVIDIA "Blackwell" RTX architecture that green team says will bring game-changing AI and neural rendering capabilities to gamers and creators,