gpu4.pics
data via LocalScore
NVIDIA RTX 4080
llamafile 0.9.2
16 GB VRAM NVIDIA 320 W TDP
running Llama 3.2 1B Instruct· Q4_K_M· 1.5B
130
tokens / sec
2,001 · Excellent
Generation
130tok/s
higher is better
Prompt / prefill
11,035tok/s
higher is better
Time to first token
179 ms
lower is better
Efficiency
1.12 tok/s/W
Electricity cost
$0.043/Mtok
Power draw
118 W
All benchmarks on this rig
| Model | Quant | Gen | Prompt | TTFT | Power | Perf/W | $/Mtok | Score |
|---|---|---|---|---|---|---|---|---|
| Llama 3.2 1B Instruct1.5B | Q4_K_M | 130 | 11,035 | 179 ms | 118 W | 1.12 tok/s/W | $0.043/Mtok | 2,001 |