TEST #1382 RESULTS
08/15/2025 - 4:34 AM
ACCELERATOR
MODEL
29.2
tokens/s
generation
1.26
sec
time to first token
1027
tokens/s
prompt
288
LocalScore
HOW YOU STACK UP
Explore All ResultsGemma 3 12b It - Q4_K - Medium
SYSTEM
CPU
Intel Xeon W-11955M CPU @ 2.60GHz (tigerlake)
RAM
125.1GB
OS
Linux
Kernel Release
6.12.10-76061203-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #202412060638~1753385872~22.04~dc2e00d SMP PREEMPT_DYNAMIC Thu J
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1118
tokens/s
28.3
tokens/s
953
ms
pp4096+tg256
1013
tokens/s
25.8
tokens/s
4.08
sec
pp2048+tg256
1090
tokens/s
28.9
tokens/s
1.91
sec
pp2048+tg768
1154
tokens/s
29.2
tokens/s
1.81
sec
pp1024+tg1024
1192
tokens/s
29.6
tokens/s
894
ms
pp1280+tg3072
1171
tokens/s
28.4
tokens/s
1.13
sec
pp384+tg1152
1126
tokens/s
30.3
tokens/s
373
ms
pp64+tg1024
984
tokens/s
31.3
tokens/s
95
ms
pp16+tg1536
393
tokens/s
30.7
tokens/s
70
ms