Home
Latest Results
Download
About
Blog
LLaMA v2
Q8_0
6.7B
params
COMPARE ACCELERATORS
1 accelerators tested
Select Accelerators
Apple M4 Pro 8P+4E+16GPU
24GB
LLaMA v2 - Q8_0
Prompt tokens/s
Generation tokens/s
Time to First Token (ms)
LocalScore
LEADERBOARD
GPU
CPU+GPU
CPU
ACCELERATOR
PROMPT
GENERATION
TTFT
LOCALSCORE
Apple M4 Pro 8P+4E+16GPU
GPU / 24GB
PROMPT
335
tokens/s
GENERATION
25.5
tokens/s
TTFT
3.69
sec
LOCALSCORE
132
