TEST #3846 RESULTS
07/02/2026 - 2:51 AM
ACCELERATOR
MODEL
94.4
tokens/s
generation
405
ms
time to first token
3189
tokens/s
prompt
906
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
13th Gen Intel Core i5-13500 (alderlake)
RAM
62.5GB
OS
Linux
Kernel Release
6.17.0-35-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #35~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue May 26 19:30:42 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3796
tokens/s
99.1
tokens/s
280
ms
pp4096+tg256
3030
tokens/s
85.3
tokens/s
1.36
sec
pp2048+tg256
3512
tokens/s
93.5
tokens/s
594
ms
pp2048+tg768
3500
tokens/s
92.3
tokens/s
596
ms
pp1024+tg1024
3775
tokens/s
95.9
tokens/s
281
ms
pp1280+tg3072
3642
tokens/s
89.9
tokens/s
362
ms
pp384+tg1152
3812
tokens/s
96.8
tokens/s
111
ms
pp64+tg1024
2791
tokens/s
99.3
tokens/s
32
ms
pp16+tg1536
841
tokens/s
98.0
tokens/s
29
ms