TEST #3875 RESULTS
07/08/2026 - 1:48 AM
ACCELERATOR
MODEL
16.8
tokens/s
generation
22.56
sec
time to first token
65
tokens/s
prompt
36
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon Platinum 8259CL CPU @ 2.50GHz (cascadelake)
RAM
15.4GB
OS
Linux
Kernel Release
6.1.166-197.305.amzn2023.x86_64
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Mon Mar 23 09:53:26 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
66
tokens/s
17.5
tokens/s
15.68
sec
pp4096+tg256
53
tokens/s
13.2
tokens/s
77.58
sec
pp2048+tg256
61
tokens/s
16.2
tokens/s
33.75
sec
pp2048+tg768
61
tokens/s
15.7
tokens/s
33.83
sec
pp1024+tg1024
66
tokens/s
17.3
tokens/s
15.65
sec
pp1280+tg3072
65
tokens/s
15.2
tokens/s
19.81
sec
pp384+tg1152
71
tokens/s
18.4
tokens/s
5.48
sec
pp64+tg1024
71
tokens/s
19.0
tokens/s
953
ms
pp16+tg1536
69
tokens/s
18.7
tokens/s
281
ms