Nvidia’s Flagship RTX 5090 GPU Could Deliver Up to 46 Percent Improvement

Nvidia GeForce RTX 50 Series Blackwell GPUs were announced at the Consumer Electronics Show (CES) 2025 earlier this month but are yet to hit the market. The RTX 5090 and RTX 5080 models will be available for purchase on January 30. Ahead of their arrival, reviewers have reportedly revealed the synthetic benchmarks for the top-of-the-line GeForce RTX 5090 GPU which hint towards a performance improvement of up to 46 percent, depending on the benchmarking platform used to run the tests.

Nvidia GeForce RTX 5090 GPU Benchmarks

In a report, VideoCardz compiled the benchmark test results. Citing conversations with several reviewers, the publication reports that gamers can expect a performance jump of approximately 20 percent with the new Nvidia GeForce RTX 5090 GPU compared to its predecessor, the RTX 4090. While these numbers are fairly accurate, the report emphasises that performance may vary significantly depending on the game, resolution, and settings.

FireStrike, which is the DirectX 11 benchmark for gaming PCs, reportedly highlighted a 36 percent performance improvement. It was evaluated in Performance, Extreme, and Ultra scenarios. Meanwhile, the DirectX 12 TimeSpy benchmark revealed a 33 percent improvement over the RTX 4090 GPU on Performance and Extreme settings.

Other reviewers also ran benchmark tests such as SpeedWay, Port Royal, and Steel Nomad and came to the conclusion that the overall performance gain could be as high as 46 percent. It should be noted that these are overall graphics scores and not the chip’s overall performance numbers. There might be discrepancy between numbers achieved in synthetic benchmarks and real-world gaming scenarios.

Nvidia GeForce RTX 5090 GPU Specifications

Nvidia’s GeForce RTX 5090 GPU is equipped with 32GB of GDDR7 VRAM (up from 24GB of GDDR6 VRAM in RTX 4090 GPU), has a peak clock speed of 2.4GHz, and gets 21,760 CUDA cores. It is built using Nvidia’s Blackwell architecture with fifth generation Tensor cores and fourth generation ray tracing cores.

Overall, the GPU features 92 billion transistors, capable of delivering 3,352 trillion operations per second (TOPS). Nvidia claims that it is the company’s most powerful GPU for consumer PCs to date.