NVIDIA GeForce RTX 5080 vs GeForce RTX 4090
NVIDIA GeForce RTX 5080
GeForce RTX 4090
NVIDIA GeForce RTX 5080 and GeForce RTX 4090 perform within a few percent of each other for gaming. Both sit in the same class, and our framerate data has each of them limited by the processor rather than the card itself, so neither holds a decisive lead. Choose on price, power draw and the features you actually use.
NVIDIA GeForce RTX 5080 vs GeForce RTX 4090 at a glance
Each axis is scored against the better of the two, so the outer edge is always the stronger part on that measure, not an absolute scale.
- Raster
- Memory
- Bandwidth
- Efficiency
- Value
NVIDIA GeForce RTX 5080 vs GeForce RTX 4090 benchmark scores
What is the difference between the NVIDIA GeForce RTX 5080 and GeForce RTX 4090?
4 specifications separate the two cards:
- Memory 16 GB GDDR7 24 GB GDDR6X
- Bandwidth 960 GB/s 1008 GB/s
- Board power 360 450
- Architecture Blackwell Ada Lovelace
Which one should you buy?
Choose the NVIDIA GeForce RTX 5080 if
- Lower power draw (360 W)
- Cheaper at launch ($999)
Choose the GeForce RTX 4090 if
- More video memory (24 GB)
How many FPS do the NVIDIA GeForce RTX 5080 and GeForce RTX 4090 get?
Estimated framerates with both parts measured against the same reference processor (a AMD Ryzen 7 9800X3D), so the graphics card is the only thing changing. Figures are ranges, not exact numbers. Measured at 1440p because at 1080p even a fast processor caps both cards and the columns come back identical.
NVIDIA GeForce RTX 5080 vs GeForce RTX 4090: full specifications and benchmarks
| Measure | NVIDIA GeForce RTX 5080 | GeForce RTX 4090 |
|---|---|---|
| Headline | ||
| G3D Mark | 35,633 | 38,890 |
| Faster than | 99% of 268 | 99% of 268 |
| Performance | ||
| 3DMark Steel Nomad | 8,858 | 9,237 |
| 3DMark Time Spy | 31,987 | 36,676 |
| Memory | ||
| Memory | 16 GB GDDR7 | 24 GB GDDR6X |
| Memory bus | 256-bit | 384-bit |
| Bandwidth | 960 GB/s | 1008 GB/s |
| Clocks and power | ||
| Boost clock | 2617 MHz | 2520 MHz |
| FP32 | 56.3 TFLOPS | 82.6 TFLOPS |
| Board power | 360 W | 450 W |
| Suggested PSU | 850 W | 850 W |
| Power connectors | 1x 16-pin (12V-2x6) | 1x 16-pin (4x 8-pin adapter) |
| Platform | ||
| Architecture | Blackwell | Ada Lovelace |
| Process | 4 nm | 5 nm |
| Price and release | ||
| Launch price | $999 | $1,599 |
| Released | 2025-01-30 | 2022-10-12 |
Which is better value?
At launch pricing, the NVIDIA GeForce RTX 5080 delivers more benchmark performance per dollar, by about 47%. Street prices move, so check current cost before deciding: a 47% value gap disappears quickly on discount.
Will the GeForce RTX 4090 bottleneck your build?
These figures hold one component constant. Your own build decides the answer: a faster graphics card buys nothing if the processor is already the limit.
How we compare graphics cards
Benchmark scores come from PassMark, 3DMark, Geekbench and Cinebench. Each of those suites is read from a single source across every part we compare, so a percentage gap here is a difference between the two chips rather than between two testers. Framerates come from our own model, built from measured GPU baselines and processor ceilings, and it declines to produce a number for hardware it does not cover rather than estimating one.