TEST #3259 RESULTS
04/09/2026 - 12:48 AM
ACCELERATOR
MODEL
14.1
tokens/s
generation
36.88
sec
time to first token
40
tokens/s
prompt
25
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
12th Gen Intel Core i5-12500T (alderlake)
RAM
15.7GB
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
41
tokens/s
15.4
tokens/s
25.15
sec
pp4096+tg256
32
tokens/s
10.7
tokens/s
127.50
sec
pp2048+tg256
37
tokens/s
13.5
tokens/s
56.05
sec
pp2048+tg768
37
tokens/s
13.3
tokens/s
55.02
sec
pp1024+tg1024
41
tokens/s
14.4
tokens/s
25.08
sec
pp1280+tg3072
40
tokens/s
12.5
tokens/s
32.25
sec
pp384+tg1152
43
tokens/s
15.3
tokens/s
9.01
sec
pp64+tg1024
46
tokens/s
16.1
tokens/s
1.44
sec
pp16+tg1536
43
tokens/s
15.7
tokens/s
429
ms