According to a Geekbench leak, the NVIDIA GeForce RTX 4060 is almost 20% faster than the RTX 3060.

NVIDIA has stayed quiet regarding the performance of the smaller Ada Lovelace AD107 GPU as it prepares to release its entry-level GeForce RTX 4060 graphics card series next week. There are rumors that this more affordable model, with an MSRP of $299, has 3,072 CUDA cores, 24 RT cores, 96 Tensor cores, 96 TMUs, and 32 ROPs. It’s believed to have an 8 GB GDDR6 memory chip with a 128-bit memory bus. Benchleaks obtained the initial batch of test results through a database leak and published them earlier today on social media. A test system using an Intel Core i5-13600K CPU, an ASUS Z790 ROG APEX motherboard, DDR5-6000 RAM, and the aforementioned GeForce card underwent two Geekbench 6 runs.

A GPU Compute test using the Vulkan API produced a score of 99419, and one using the OpenCL API produced a score of 105630. Since this is just one sample, there may be differences when additional units are examined in Geekbench before the launch on June 29. The RTX 4060 is roughly 12% faster (in Vulkan) than the RTX 3060, which it directly replaces. With its Open CL performance, where it provides a nearly 20% improvement over the previous card, the gap grows. Over the RTX 4060, the RTX 3060 Ti performs about 3-5% quicker. We anticipate seeing genuine in-game benchmarks soon.