TEST #1903 RESULTS
12/03/2025 - 8:14 PM
ACCELERATOR
MODEL
40.9
tokens/s
generation
652
ms
time to first token
2108
tokens/s
prompt
510
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i9-14900HX (alderlake)
RAM
95.7GB
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
2577
tokens/s
41.2
tokens/s
421
ms
pp4096+tg256
1769
tokens/s
39.5
tokens/s
2.34
sec
pp2048+tg256
2257
tokens/s
41.4
tokens/s
932
ms
pp2048+tg768
2255
tokens/s
41.1
tokens/s
932
ms
pp1024+tg1024
2590
tokens/s
41.0
tokens/s
420
ms
pp1280+tg3072
2517
tokens/s
40.6
tokens/s
533
ms
pp384+tg1152
3064
tokens/s
39.6
tokens/s
149
ms
pp64+tg1024
1636
tokens/s
41.3
tokens/s
65
ms
pp16+tg1536
311
tokens/s
42.3
tokens/s
75
ms