gpu4.pics
data via LocalScore
NVIDIA RTX 4090
24 GB VRAM NVIDIA 450 W TDP
running Meta Llama 3.1 8B Instruct· Q4_K_M
90.3
tokens / sec
1,439 · Excellent #4 of 370
Generation
90.3tok/s
higher is better
Prompt / prefill
6,661tok/s
higher is better
Time to first token
205 ms
lower is better
Composite score
1,439
Model size
8B
Runtime
—
All benchmarks on this rig
| Model | Quant | Gen | Prompt | TTFT | Power | Perf/W | $/Mtok | Score |
|---|---|---|---|---|---|---|---|---|
| Meta Llama 3.1 8B Instruct | Q4_K_M | 90.3 | 6,661 | 205 ms | — | — | — | 1,439 |