gpu4.pics
data via LocalScore
NVIDIA GeForce RTX 3070 Ti
8 GB VRAM NVIDIA
running Meta Llama 3.1 8B Instruct· Q4_K_M
83.2
tokens / sec
738 · Strong #32 of 370
Generation
83.2tok/s
higher is better
Prompt / prefill
2,509tok/s
higher is better
Time to first token
520 ms
lower is better
Composite score
738
Model size
8B
Runtime
gpu4.pics/r/ls-nvidia-geforce-rtx-3070-ti-meta-llama-31-8b-instruct-q4-k-m

All benchmarks on this rig

ModelQuantGenPromptTTFTPowerPerf/W$/MtokScore
Meta Llama 3.1 8B InstructQ4_K_M83.22,509520 ms738

How it stacks up — Meta Llama 3.1 8B Instruct

NVIDIA GeForce RTX 3070 Ti
83.2
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