TEST #2897 RESULTS
02/24/2026 - 9:53 PM
ACCELERATOR
MODEL
12.6
tokens/s
generation
55.04
sec
time to first token
24
tokens/s
prompt
18
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Virtual CPU @ 2.99GHz (bdver2)
RAM
15.4GB
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
25
tokens/s
14.0
tokens/s
40.89
sec
pp4096+tg256
23
tokens/s
11.3
tokens/s
175.80
sec
pp2048+tg256
25
tokens/s
12.1
tokens/s
82.87
sec
pp2048+tg768
24
tokens/s
11.4
tokens/s
84.19
sec
pp1024+tg1024
24
tokens/s
13.1
tokens/s
42.32
sec
pp1280+tg3072
25
tokens/s
11.7
tokens/s
51.04
sec
pp384+tg1152
26
tokens/s
13.4
tokens/s
14.84
sec
pp64+tg1024
26
tokens/s
13.2
tokens/s
2.55
sec
pp16+tg1536
21
tokens/s
12.8
tokens/s
816
ms