TEST #2083 RESULTS
01/03/2026 - 4:22 PM
ACCELERATOR
MODEL
22.5
tokens/s
generation
1.39
sec
time to first token
1012
tokens/s
prompt
254
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core Ultra 7 265H (arrowlake)
RAM
30.3GB
OS
Linux
Kernel Release
6.17.0-8-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #8-Ubuntu SMP PREEMPT_DYNAMIC Fri Nov 14 21:44:46 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1099
tokens/s
23.6
tokens/s
955
ms
pp4096+tg256
878
tokens/s
22.0
tokens/s
4.71
sec
pp2048+tg256
1007
tokens/s
22.4
tokens/s
2.06
sec
pp2048+tg768
1002
tokens/s
22.3
tokens/s
2.07
sec
pp1024+tg1024
1094
tokens/s
22.6
tokens/s
968
ms
pp1280+tg3072
1063
tokens/s
21.5
tokens/s
1.23
sec
pp384+tg1152
1156
tokens/s
22.5
tokens/s
361
ms
pp64+tg1024
1276
tokens/s
22.7
tokens/s
74
ms
pp16+tg1536
535
tokens/s
22.6
tokens/s
53
ms