gpu4.pics
data via LocalScore
NVIDIA RTX PRO 6000 Blackwell Workstation Edition
95 GB VRAM NVIDIA
running Qwen2.5 14B Instruct· Q4_K_M· 14.8B
70.8
tokens / sec
1,080 · Excellent #2 of 373
Generation
70.8tok/s
higher is better
Prompt / prefill
4,761tok/s
higher is better
Time to first token
281 ms
lower is better
Composite score
1,080
Model size
14.8B
Runtime
—
All benchmarks on this rig
| Model | Quant | Gen | Prompt | TTFT | Power | Perf/W | $/Mtok | Score |
|---|---|---|---|---|---|---|---|---|
| Qwen2.5 14B Instruct14.8B | Q4_K_M | 70.8 | 4,761 | 281 ms | — | — | — | 1,080 |