As NVIDIA’s new GeForce RTX 5090 is gearing up for launch, speculations naturally follow on how good the card’s performance will be in its bare form. Recently, several Geekbench benchmark entries shows just that – while measured purely on synthetic performance, such numbers can provide a basic insight on what’s to come.
RTX 5090 Spotted In Benchmarks
These benchmark entries likely originated from a reviewer testing the GPU, with one entry listed the system powered by Intel Core i9-12900K with an ASUS ROG Z690 motherboard paired with DDR4-3600 RAM. As per Videocardz, the three numbers found on the database shows 359,742, 359,557, and 331,525 points – compared to RTX 4090 in Vulkan tests (around 260K points), the card is faster with a margin of at least 26% but no more than 37%.
OpenCL is a different story, however. All three entries shows points ranging from 338K to 367K, while the RTX 4090 scored 317K; meaning, the new card is measured to only outperform the outgoing GPU by the range of 6.7% to 15%. Either way, these are synthetic benchmarks testing a specific workload, and games are naturally requiring a mix of various components which makes this data difficult to compare directly.
Interestingly, the GPU is also spotted in Blender’s Open Data benchmark database, and this may be more representative of the card’s performance since it focuses on 3D rendering performance. Using NVIDIA’s propriety OptiX engine, the numbers indicate the RTX 5090 is around 37% faster than RTX 4090’s average score (which rules out any overclocked results that may skew the numbers). In this case, it does seem to align with numbers estimated through NVIDIA’s first-party gaming benchmarks.
In any case, we should be able to see the review of RTX 5090 pretty soon as Team Green will be launching this new GPU by the end of this month, as announced in CES 2025.
Pokdepinion: Would it be a worthwhile upgrade for RTX 4090 owners?
