TEST #1829 RESULTS
11/19/2025 - 6:58 PM
ACCELERATOR
MODEL
92.1
tokens/s
generation
406
ms
time to first token
3156
tokens/s
prompt
894
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 5600GT with Radeon Graphics (znver3)
RAM
13.5GB
OS
Linux
Kernel Release
6.14.0-35-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #35~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Tue Oct 14 13:55:17 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3805
tokens/s
96.2
tokens/s
280
ms
pp4096+tg256
3022
tokens/s
83.4
tokens/s
1.37
sec
pp2048+tg256
3517
tokens/s
90.7
tokens/s
593
ms
pp2048+tg768
3497
tokens/s
89.3
tokens/s
596
ms
pp1024+tg1024
3796
tokens/s
93.6
tokens/s
280
ms
pp1280+tg3072
3681
tokens/s
88.2
tokens/s
358
ms
pp384+tg1152
3920
tokens/s
95.1
tokens/s
109
ms
pp64+tg1024
2643
tokens/s
96.6
tokens/s
34
ms
pp16+tg1536
527
tokens/s
95.8
tokens/s
40
ms