TEST #379 RESULTS
04/06/2025 - 3:05 PM
MODEL
3.3
tokens/s
generation
54.29
sec
time to first token
24
tokens/s
prompt
11
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
RAM
7.4GB
OS
Linux
Kernel Release
4.14.116
Architecture
aarch64
Version
Cosmopolitan 3.9.7 MODE=aarch64; #1 SMP PREEMPT Tue Aug 16 18:48:58 CST 2022
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
29
tokens/s
4.1
tokens/s
35.39
sec
pp4096+tg256
24
tokens/s
2.7
tokens/s
172.36
sec
pp2048+tg256
24
tokens/s
3.0
tokens/s
86.19
sec
pp2048+tg768
23
tokens/s
2.9
tokens/s
89.67
sec
pp1024+tg1024
25
tokens/s
3.3
tokens/s
41.19
sec
pp1280+tg3072
29
tokens/s
3.3
tokens/s
44.15
sec
pp384+tg1152
27
tokens/s
3.4
tokens/s
14.33
sec
pp64+tg1024
20
tokens/s
3.3
tokens/s
3.57
sec
pp16+tg1536
11
tokens/s
3.7
tokens/s
1.78
sec