TEST #1708 RESULTS
10/28/2025 - 3:37 AM
ACCELERATOR
MODEL
46.9
tokens/s
generation
228
ms
time to first token
5573
tokens/s
prompt
1047
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 9354 32-Core Processor (znver4)
RAM
377.4GB
OS
Linux
Kernel Release
6.8.0-85-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #85-Ubuntu SMP PREEMPT_DYNAMIC Thu Sep 18 15:26:59 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
7263
tokens/s
50.6
tokens/s
161
ms
pp4096+tg256
6096
tokens/s
37.8
tokens/s
697
ms
pp2048+tg256
6891
tokens/s
48.3
tokens/s
318
ms
pp2048+tg768
6893
tokens/s
45.0
tokens/s
318
ms
pp1024+tg1024
7288
tokens/s
48.7
tokens/s
161
ms
pp1280+tg3072
6688
tokens/s
42.3
tokens/s
212
ms
pp384+tg1152
6676
tokens/s
49.3
tokens/s
77
ms
pp64+tg1024
1819
tokens/s
50.0
tokens/s
55
ms
pp16+tg1536
542
tokens/s
49.7
tokens/s
49
ms