TEST #3090 RESULTS
03/15/2026 - 2:01 PM
ACCELERATOR
MODEL
48.1
tokens/s
generation
391
ms
time to first token
3235
tokens/s
prompt
736
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 7800X3D 8-Core Processor (znver4)
RAM
61.9GB
OS
Linux
Kernel Release
6.18.7-76061807-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #202601231045~1769703228~24.04~cb87b5b SMP PREEMPT_DYNAMIC Thu J
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
4200
tokens/s
52.9
tokens/s
263
ms
pp4096+tg256
3237
tokens/s
39.0
tokens/s
1.29
sec
pp2048+tg256
3822
tokens/s
46.4
tokens/s
557
ms
pp2048+tg768
3838
tokens/s
45.1
tokens/s
555
ms
pp1024+tg1024
4225
tokens/s
48.9
tokens/s
261
ms
pp1280+tg3072
4030
tokens/s
43.5
tokens/s
337
ms
pp384+tg1152
3870
tokens/s
51.6
tokens/s
119
ms
pp64+tg1024
1397
tokens/s
53.4
tokens/s
68
ms
pp16+tg1536
492
tokens/s
52.4
tokens/s
63
ms