GPU comparison
NVIDIA RTX 3060 (12GB) vs NVIDIA RTX 4080 (16GB) for local LLMs
How the two stack up for running open-source LLMs locally — memory, bandwidth, price, and how many of the 18 tracked models each one can run.
| NVIDIA RTX 3060 (12GB) | NVIDIA RTX 4080 (16GB) | |
|---|---|---|
| Memory | 12.0 GB | 16.0 GB |
| Bandwidth | 360 GB/s | 717 GB/s |
| Price (approx) | $289 | $999 |
| LLMs it runs | 8 of 18 | 12 of 18 |
| Best model it runs | Gemma 4 12B · 31–46 tok/s | Gemma 4 26B-A4B (MoE) · 107–161 tok/s |
The NVIDIA RTX 4080 (16GB) runs 4 more of the tracked models (12 vs 8), thanks to its 16.0 GB of memory. The NVIDIA RTX 4080 (16GB) has more memory bandwidth (717 GB/s), so it generates tokens faster at the same model and quant.