According to a recent leak, the NVIDIA RTX 5080 is slower than the RTX 4090.

The performance capabilities of NVIDIA’s future RTX 5080 GPU have been made public by a set of recently leaked benchmarks. The GPU, which is slated to arrive on January 30 along with the RTX 5090, was observed on Geekbench running OpenCL and Vulkan benchmark tests. Based on its performance, it may not rank among the top graphics cards. The MSI-branded RTX 5080, model MS-7E62, was the device under test. AMD’s Ryzen 7 9800X3D chip, which is regarded by many as one of the top CPUs for gaming, was installed in this setup. Additionally, it came with 32 GB of DDR5-6000 memory and an MSI MPG 850 Edge TI Wi-Fi motherboard.

According to benchmark findings, the RTX 5080 received 256,138 points in OpenCL tests and 261,836 points in Vulkan testing. The RTX 5080 features a slight 6.7% increase in OpenCL performance and a 22% increase in Vulkan performance over its predecessor, the RTX 4080. The GPU received a median score of 9,063.77 on the Blender Open Data platform, according to Reddit user TruthPhoenixV. This score outperforms the RTX 4080 Super by 8.2% and the RTX 4080 by 9.4%. The top-tier RTX 4090 from the current generation may still perform better than the RTX 5080 despite these enhancements. NVIDIA’s 80-class GPUs have historically outperformed the 90-class GPUs from the preceding generation, but our preliminary results imply that the RTX 5080 may not follow this pattern.

Using NVIDIA’s most recent Blackwell architecture, the RTX 5080 has 10,752 CUDA cores distributed among 84 Streaming Multiprocessors (SMs), as opposed to the RTX 4080’s 9,728 cores. It features a 256-bit bus and 16 GB of GDDR7 memory. According to NVIDIA, its Tensor Cores can achieve 1,801 TOPS in AI performance, while its RT Cores can achieve 171 TeraFLOPS in ray tracing performance.

However, it’s crucial to remember that these benchmark findings haven’t been thoroughly confirmed, so we shouldn’t draw any conclusions until the review embargo has ended.