Side-by-side simulation · recommended settings
RTX 4080 SUPER vs RTX 5080: A 16GB local LLM upgrade?
If your model already runs well on the 4080 SUPER, the 5080 is a speed upgrade. To expand model capacity, compare 24GB and 32GB cards too.
Choose devices and a model
Expand comparison settings
Same 117output tokens. Only evidence-qualified acceleration is selected automatically; runtime formats are shown on each device.
Left
ReadyRTX 4080 SUPER 16GB
Press Start above to generate the same answer on both devices.
Right
ReadyRTX 5080 16GB
Press Start above to generate the same answer on both devices.
With these settings, RTX 5080 16GB finishes the full answer About 1.2× faster.
Compare from input processing to completion of 117 token response.
Which one fits my needs?
Answer completion
RTX 5080 16GB
At the estimate midpoint, it finishes the same answer about 1.2× faster.
Prompt processing
RTX 5080 16GB
The estimated input-processing midpoint is higher at 2,148 tok/s.
Image and video generation
Check the media model
LLM speed does not tell you media-generation performance. Choose the image or video model and resolution separately.
Runtime
Settings for your hardware
Strata · Ada Lovelace (CC 8.9) · CUDA · MTP · NVMe PLE · single GPU / Strata · Blackwell (CC 12.0) · CUDA · MTP · NVMe PLE · single GPU. When runtime paths differ, the speed gap also includes software and model-format differences.
Estimated power draw
RTX 4080 SUPER 16GB
Estimated LLM power draw is lower at about 324 W. Noise and physical size need a separate check.
Local pricing
RTX 5080 16GB costs about 45 × KRW 10,000 more at the reference-price midpoint, with an estimated answer time about 18% shorter.
Before choosing hardware
| Measure | RTX 4080 SUPER 16GB | RTX 5080 16GB |
|---|---|---|
| Available fast memory | 15GB | 15GB |
| Memory bandwidth | 736GB/s | 960GB/s |
| Estimated total time | 3.8 s | 3.1 s |
| Estimated sustained LLM load | About 324W | About 352W |
| Korean price range (KRW) | 175–205 × KRW 10,000 | 231–239 × KRW 10,000 |
Estimated ranges for one user. Multi-GPU and multi-device results depend on the runtime and interconnect.
Does more bandwidth help your conversations?
Generating an answer token by token makes fast weight reads important. Check bandwidth, then run decode mode. If you often process long input, compare full mode too. Looking at both stages gives a better account of workload time than judging from one metric.
Distinguish Blackwell files from your existing ones
Fast quantized paths on the 5080 require the runtime's matching files and packages. Your 4080 SUPER's GGUF setup may differ from the new recommendation. Update the existing card's supported path too, so you can distinguish a hardware upgrade from a software improvement.
Use a larger-memory comparison for workloads beyond 16GB
If loading still fails after reducing context, or your model requires a smaller quantization, compare increasing capacity. Moving from a 4080 SUPER to a 5080 solves a different problem from moving to a 5090. Related comparisons below include larger-VRAM options.