TEST #3527 RESULTS
05/14/2026 - 12:32 PM
ACCELERATOR
MODEL
1.3
tokens/s
generation
31.48
sec
time to first token
45
tokens/s
prompt
12
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2650 v4 @ 2.20GHz (broadwell)
RAM
63.9GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
42
tokens/s
1.3
tokens/s
25.00
sec
pp4096+tg256
35
tokens/s
1.0
tokens/s
119.36
sec
pp2048+tg256
50
tokens/s
1.2
tokens/s
41.98
sec
pp2048+tg768
49
tokens/s
1.2
tokens/s
42.36
sec
pp1024+tg1024
55
tokens/s
1.5
tokens/s
19.24
sec
pp1280+tg3072
55
tokens/s
1.3
tokens/s
24.10
sec
pp384+tg1152
55
tokens/s
1.4
tokens/s
7.71
sec
pp64+tg1024
52
tokens/s
1.5
tokens/s
1.91
sec
pp16+tg1536
16
tokens/s
1.4
tokens/s
1.66
sec