DGX Spark vs RTX 3090
For running local LLMs · 15 reports across 4 models
Tokens per second by model
| Model | DGX Spark | RTX 3090 |
|---|---|---|
| Qwen3.6up to 35B | — | 104.0n=10 |
| DeepSeek V3 | —n=1 | —n=1 |
| Qwen2.5 | — | 70.0n=2 |
| GLM-5.1 | —n=1 | — |
For running local LLMs · 15 reports across 4 models
| Model | DGX Spark | RTX 3090 |
|---|---|---|
| Qwen3.6up to 35B | — | 104.0n=10 |
| DeepSeek V3 | —n=1 | —n=1 |
| Qwen2.5 | — | 70.0n=2 |
| GLM-5.1 | —n=1 | — |