RTX 3090 vs RTX Pro 6000 Blackwell
For running local LLMs · 120 reports across 15 models
Highest reported generation speed by model family
These maxima come from different community setups. Model size, quantization, GPU count, offloading and concurrency may differ. They do not establish which GPU is faster for the same workload. Check the comparison conditions →
| Model | RTX 3090 | RTX Pro 6000 Blackwell |
|---|---|---|
| Qwen3.8up to 2400B | 382.0n=56 | 177.0n=16 |
| Qwen3.6up to 35B | 170.7n=13 | 600.0n=5 |
| DeepSeek V4 Flashup to 285B | 60.0n=9 | 45.0n=4 |
| Museup to 30B | 94.0n=1 | 57.0n=2 |
| GLM-5.2up to 744B | 7.3n=2 | — |
| GLM-5.3up to 321B | 55.0n=2 | — |
| Qwen2.5up to 32B | 70.0n=2 | — |
| DeepSeek R1 | 2.0n=1 | — |
| DeepSeek V3 | —n=1 | — |
| DeepSeek V4up to 284B | 15.0n=1 | — |
| GLM-4.5-Airup to 106B | —n=1 | — |
| Llama 3.2up to 1B | — | —n=1 |
| Mimo 2.5 | — | —n=1 |
| Qwen3up to 235B | 7.5n=1 | — |
| Tencent-HY3up to 295B | — | 57.4n=1 |