TEST #806 RESULTS
05/03/2025 - 11:08 PM
ACCELERATOR
MODEL
39.7
tokens/s
generation
912
ms
time to first token
1376
tokens/s
prompt
391
LocalScore
HOW YOU STACK UP
Explore All ResultsGemma 3 12b - Q4_0
SYSTEM
CPU
Intel Xeon W-3235 CPU @ 3.30GHz (cascadelake)
RAM
41GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1618
tokens/s
42.2
tokens/s
659
ms
pp4096+tg256
1428
tokens/s
33.9
tokens/s
2.90
sec
pp2048+tg256
1531
tokens/s
39.0
tokens/s
1.36
sec
pp2048+tg768
1525
tokens/s
38.0
tokens/s
1.37
sec
pp1024+tg1024
1593
tokens/s
40.3
tokens/s
668
ms
pp1280+tg3072
1569
tokens/s
36.7
tokens/s
841
ms
pp384+tg1152
1671
tokens/s
41.9
tokens/s
253
ms
pp64+tg1024
1120
tokens/s
43.4
tokens/s
79
ms
pp16+tg1536
327
tokens/s
42.3
tokens/s
71
ms