gpu4.pics
data via LocalScore
NVIDIA H100
80 GB VRAM NVIDIA 700 W TDP
running Meta Llama 3.1 8B Instruct· Q4_K_M
120
tokens / sec
1,382 · Excellent #5 of 370
Generation
120tok/s
higher is better
Prompt / prefill
5,202tok/s
higher is better
Time to first token
236 ms
lower is better
Composite score
1,382
Model size
8B
Runtime
gpu4.pics/r/ls-nvidia-h100-meta-llama-31-8b-instruct-q4-k-m

All benchmarks on this rig

ModelQuantGenPromptTTFTPowerPerf/W$/MtokScore
Meta Llama 3.1 8B InstructQ4_K_M1205,202236 ms1,382

How it stacks up — Meta Llama 3.1 8B Instruct

NVIDIA H100
120
NVIDIA A100 80GB
110
NVIDIA RTX 3090 Ti
110
NVIDIA RTX PRO 6000 Blackwell Workstation Edition
103
NVIDIA RTX 3090
103
NVIDIA RTX A6000
90.5