GPU comparison · 1080p benchmark
Compare RTX 4070 12GB and RTX 5080 16GB using GPUsed's recorded model-level benchmark FPS. See the percentage difference, then open the full comparison tool if you want to choose another card.
RTX 5080 16GB recorded 538 FPS against 285 FPS for RTX 4070 12GB. These are continuous 1080p benchmark averages, not predicted FPS in any particular game.
RTX 4070 12GB
285Benchmark FPS at 1080p- Benchmark score
- 36.4 / 100
- Graphics memory
- 12GB
- Released
- April 2023
RTX 5080 16GB
538Benchmark FPS at 1080p- Benchmark score
- 68.6 / 100
- Graphics memory
- 16GB
- Released
- January 2025
Change either GPU in the comparison tool
What does this comparison mean?
The 5080 substantially outruns the 4070 and offers 16GB instead of 12GB. This is a higher-tier move, not a direct successor comparison.
RTX 4070 12GB has 12GB and RTX 5080 16GB has 16GB of graphics memory. Memory capacity, board dimensions, power connectors and the price of a particular used card are separate from this benchmark. Compare the exact item before deciding whether to upgrade.
The two GPUs in context
RTX 4070 12GB
The RTX 4070 offered Ada efficiency and 12GB memory in a more approachable 70-class card. Its restrained power draw gave it a practical identity, although the later 4070 Super made the original position less clear-cut.
RTX 5080 16GB
The RTX 5080 is a capable 4K Blackwell card, yet its native-rendering advance over the 4080 Super was modest for a new generation. It remains a high-end option, with a legacy still too young to judge firmly.
Release dates: RTX 4070 12GB source · RTX 5080 16GB source.
How the figures are compared
Both FPS figures come from the owner-supplied GPUsed benchmark dataset. GPUsed reports a continuous FurMark run at 1080p for its physical-card checks. The percentage above compares the higher model average with the lower one; the score scales every model against the fastest GPU in the current dataset, which scores 100. A benchmark result does not predict gaming FPS, and the percentage may change with the benchmark dataset.