TEST #3766 RESULTS
06/20/2026 - 3:24 AM
ACCELERATOR
MODEL
53.2
tokens/s
generation
585
ms
time to first token
2340
tokens/s
prompt
597
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 5 3600 6-Core Processor (znver2)
RAM
15.1GB
OS
Linux
Kernel Release
7.0.0-22-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #22-Ubuntu SMP PREEMPT_DYNAMIC Mon May 25 15:54:34 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2836
tokens/s
58.5
tokens/s
378
ms
pp4096+tg256
2019
tokens/s
39.4
tokens/s
2.05
sec
pp2048+tg256
2485
tokens/s
50.0
tokens/s
844
ms
pp2048+tg768
2485
tokens/s
48.3
tokens/s
844
ms
pp1024+tg1024
2836
tokens/s
54.5
tokens/s
378
ms
pp1280+tg3072
2715
tokens/s
45.9
tokens/s
489
ms
pp384+tg1152
3264
tokens/s
59.0
tokens/s
133
ms
pp64+tg1024
2233
tokens/s
62.7
tokens/s
43
ms
pp16+tg1536
191
tokens/s
60.6
tokens/s
99
ms