TEST #3473 RESULTS
05/08/2026 - 5:13 AM
ACCELERATOR
MODEL
92.4
tokens/s
generation
453
ms
time to first token
2796
tokens/s
prompt
829
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 5600G with Radeon Graphics (znver3)
RAM
7.2GB
OS
Linux
Kernel Release
6.12.85+deb13-amd64
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Debian 6.12.85-1 (2026-04-30)
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2142
tokens/s
94.1
tokens/s
489
ms
pp4096+tg256
2786
tokens/s
83.5
tokens/s
1.48
sec
pp2048+tg256
3361
tokens/s
91.4
tokens/s
620
ms
pp2048+tg768
3369
tokens/s
90.6
tokens/s
619
ms
pp1024+tg1024
3633
tokens/s
93.8
tokens/s
292
ms
pp1280+tg3072
3537
tokens/s
88.7
tokens/s
372
ms
pp384+tg1152
3590
tokens/s
95.0
tokens/s
117
ms
pp64+tg1024
2293
tokens/s
97.8
tokens/s
38
ms
pp16+tg1536
453
tokens/s
96.8
tokens/s
45
ms