TEST #3254 RESULTS
04/07/2026 - 8:12 PM
ACCELERATOR
MODEL
15.2
tokens/s
generation
45.47
sec
time to first token
31
tokens/s
prompt
22
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-3720QM CPU @ 2.60GHz (ivybridge)
RAM
15GB
OS
Linux
Kernel Release
6.14.0-37-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #37-Ubuntu SMP PREEMPT_DYNAMIC Fri Nov 14 22:10:32 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
38
tokens/s
17.9
tokens/s
26.70
sec
pp4096+tg256
29
tokens/s
12.9
tokens/s
142.90
sec
pp2048+tg256
30
tokens/s
13.8
tokens/s
69.00
sec
pp2048+tg768
28
tokens/s
13.4
tokens/s
72.48
sec
pp1024+tg1024
27
tokens/s
14.6
tokens/s
37.93
sec
pp1280+tg3072
28
tokens/s
13.2
tokens/s
45.08
sec
pp384+tg1152
30
tokens/s
16.3
tokens/s
12.79
sec
pp64+tg1024
36
tokens/s
17.3
tokens/s
1.86
sec
pp16+tg1536
35
tokens/s
17.5
tokens/s
499
ms