TEST #3291 RESULTS
04/11/2026 - 3:52 PM
ACCELERATOR
MODEL
24.0
tokens/s
generation
410
ms
time to first token
3083
tokens/s
prompt
565
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
31.1GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3911
tokens/s
24.9
tokens/s
302
ms
pp4096+tg256
3604
tokens/s
23.8
tokens/s
1.18
sec
pp2048+tg256
3900
tokens/s
24.1
tokens/s
567
ms
pp2048+tg768
3882
tokens/s
24.1
tokens/s
570
ms
pp1024+tg1024
4021
tokens/s
24.1
tokens/s
296
ms
pp1280+tg3072
3592
tokens/s
23.9
tokens/s
397
ms
pp384+tg1152
3707
tokens/s
23.7
tokens/s
152
ms
pp64+tg1024
891
tokens/s
23.6
tokens/s
112
ms
pp16+tg1536
237
tokens/s
23.6
tokens/s
115
ms