GPU comparison
NVIDIA RTX 4070 (12GB) vs NVIDIA RTX 5070 Ti (16GB) for local LLMs
How the two stack up for running open-source LLMs locally — memory, bandwidth, price, and how many of the 34 tracked models each one can run.
| NVIDIA RTX 4070 (12GB) | NVIDIA RTX 5070 Ti (16GB) | |
|---|---|---|
| Memory | 12.0 GB | 16.0 GB |
| Bandwidth | 504 GB/s | 896 GB/s |
| Price (approx) | $549 | $949 |
| LLMs it runs | 13 of 34 | 18 of 34 |
| Best model it runs | Devstral Small 2 24B · 35–53 tok/s | Gemma 4 26B-A4B (MoE) · 161–241 tok/s |
The NVIDIA RTX 5070 Ti (16GB) runs 5 more of the tracked models (18 vs 13), thanks to its 16.0 GB of memory. The NVIDIA RTX 5070 Ti (16GB) has more memory bandwidth (896 GB/s), so it generates tokens faster at the same model and quant.