TEST #3346 RESULTS
04/18/2026 - 7:54 AM
ACCELERATOR
MODEL
44.9
tokens/s
generation
1.29
sec
time to first token
1002
tokens/s
prompt
327
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2673 v3 @ 2.40GHz (haswell)
RAM
31.9GB
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
1193
tokens/s
45.3
tokens/s
881
ms
pp4096+tg256
936
tokens/s
30.8
tokens/s
4.41
sec
pp2048+tg256
1092
tokens/s
41.5
tokens/s
1.90
sec
pp2048+tg768
1093
tokens/s
39.7
tokens/s
1.90
sec
pp1024+tg1024
1194
tokens/s
46.5
tokens/s
878
ms
pp1280+tg3072
1150
tokens/s
37.3
tokens/s
1.13
sec
pp384+tg1152
1218
tokens/s
52.1
tokens/s
333
ms
pp64+tg1024
871
tokens/s
56.9
tokens/s
89
ms
pp16+tg1536
270
tokens/s
54.4
tokens/s
74
ms