TEST #3334 RESULTS
04/16/2026 - 5:32 PM
ACCELERATOR
MODEL
78.7
tokens/s
generation
724
ms
time to first token
1874
tokens/s
prompt
589
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-5820K CPU @ 3.30GHz (haswell)
RAM
15.5GB
OS
Linux
Kernel Release
5.15.0-174-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #184-Ubuntu SMP Fri Mar 13 18:41:50 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2173
tokens/s
86.1
tokens/s
483
ms
pp4096+tg256
1625
tokens/s
52.9
tokens/s
2.54
sec
pp2048+tg256
1958
tokens/s
71.8
tokens/s
1.06
sec
pp2048+tg768
1952
tokens/s
68.7
tokens/s
1.06
sec
pp1024+tg1024
2156
tokens/s
80.6
tokens/s
486
ms
pp1280+tg3072
2072
tokens/s
64.3
tokens/s
630
ms
pp384+tg1152
2280
tokens/s
90.8
tokens/s
178
ms
pp64+tg1024
1822
tokens/s
99.0
tokens/s
44
ms
pp16+tg1536
826
tokens/s
94.4
tokens/s
28
ms