RTX 3090 vs RTX 4070
For running local LLMs · 14 reports across 4 models
Tokens per second by model
| Model | RTX 3090 | RTX 4070 |
|---|---|---|
| Qwen3.6up to 35B | 104.0n=10 | — |
| Qwen2.5 | 70.0n=2 | — |
| DeepSeek V3 | —n=1 | — |
| Gemma 4up to 31B | — | 55.0n=1 |
For running local LLMs · 14 reports across 4 models
| Model | RTX 3090 | RTX 4070 |
|---|---|---|
| Qwen3.6up to 35B | 104.0n=10 | — |
| Qwen2.5 | 70.0n=2 | — |
| DeepSeek V3 | —n=1 | — |
| Gemma 4up to 31B | — | 55.0n=1 |