RTX PRO 6000 Blackwell x4
9recorded runs
231.6best tok/s
9sourced
All benchmarks Compare RTX PRO 6000 Blackwell x4 →
| GPU | Model | Quant | Backend | Ctx | tok/s | prompt tok/s | VRAM GB | Source |
|---|---|---|---|---|---|---|---|---|
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 65536 | 231.6 ✓ | 7346.3 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 2048 | 226.8 ✓ | 7010.3 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 400000 | 220.2 ✓ | 6081.4 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 500000 | 212.4 ✓ | 5844.5 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 32768 | 211.5 ✓ | 7438.7 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 131072 | 203.3 ✓ | 7064.6 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 200000 | 202.2 ✓ | 6802.6 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 512 | 200.9 ✓ | 1943.0 | — | src |
| RTX PRO 6000 Blackwell x4 · 384 GB | DeepSeek-V4.1-Flash | FP4+FP8 native | sglang lmsysorg/sglang c4ca6511 + nvme/DDR5… | 8192 | 196.4 ✓ | 7337.0 | — | src |