TEST #1323 RESULTS
08/07/2025 - 5:59 AM
ACCELERATOR
MODEL
27.9
tokens/s
generation
1.95
sec
time to first token
682
tokens/s
prompt
214
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 3B Instruct - Q6_K
SYSTEM
CPU
Intel Core i5-6600K CPU @ 3.50GHz (skylake)
RAM
15.1GB
OS
Linux
Kernel Release
6.14.0-27-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #27-Ubuntu SMP PREEMPT_DYNAMIC Tue Jul 22 17:01:58 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
775
tokens/s
31.2
tokens/s
1.35
sec
pp4096+tg256
632
tokens/s
19.9
tokens/s
6.53
sec
pp2048+tg256
716
tokens/s
26.1
tokens/s
2.90
sec
pp2048+tg768
712
tokens/s
25.1
tokens/s
2.91
sec
pp1024+tg1024
763
tokens/s
28.5
tokens/s
1.37
sec
pp1280+tg3072
742
tokens/s
23.6
tokens/s
1.76
sec
pp384+tg1152
794
tokens/s
31.2
tokens/s
513
ms
pp64+tg1024
687
tokens/s
33.3
tokens/s
121
ms
pp16+tg1536
314
tokens/s
32.1
tokens/s
78
ms