GPU comparison
NVIDIA RTX 4070 (12GB) vs NVIDIA RTX 5080 (16GB) for local LLMs
How the two stack up for running open-source LLMs locally — memory, bandwidth, price, and how many of the 64 tracked models each one can run.
| NVIDIA RTX 4070 (12GB) | NVIDIA RTX 5080 (16GB) | |
|---|---|---|
| Memory | 12.0 GB | 16.0 GB |
| Bandwidth | 504 GB/s | 960 GB/s |
| Price (approx) | $549 | $1,399 |
| LLMs it runs | 27 of 64 | 45 of 64 |
| Best model it runs | Devstral Small 2 24B · 35–53 tok/s | Qwen3.8 27B · 54–82 tok/s |
The NVIDIA RTX 5080 (16GB) runs 18 more of the tracked models (45 vs 27), thanks to its 16.0 GB of memory. The NVIDIA RTX 5080 (16GB) has more memory bandwidth (960 GB/s), so it generates tokens faster at the same model and quant.