TEST #3834 RESULTS
07/01/2026 - 4:47 AM
ACCELERATOR
MODEL
61.2
tokens/s
generation
533
ms
time to first token
2471
tokens/s
prompt
657
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
12th Gen Intel Core i7-12700K (alderlake)
RAM
31.7GB
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
2960
tokens/s
62.0
tokens/s
362
ms
pp4096+tg256
2271
tokens/s
56.2
tokens/s
1.82
sec
pp2048+tg256
2682
tokens/s
61.4
tokens/s
780
ms
pp2048+tg768
2676
tokens/s
61.3
tokens/s
782
ms
pp1024+tg1024
2953
tokens/s
62.5
tokens/s
363
ms
pp1280+tg3072
2839
tokens/s
59.5
tokens/s
467
ms
pp384+tg1152
3287
tokens/s
62.7
tokens/s
133
ms
pp64+tg1024
1907
tokens/s
62.8
tokens/s
50
ms
pp16+tg1536
669
tokens/s
62.8
tokens/s
40
ms