TEST #834 RESULTS
05/06/2025 - 8:29 AM
ACCELERATOR
MODEL
22.0
tokens/s
generation
1.95
sec
time to first token
647
tokens/s
prompt
194
LocalScore
HOW YOU STACK UP
Explore All ResultsGemma 3 27b - 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
761
tokens/s
23.4
tokens/s
1.39
sec
pp4096+tg256
669
tokens/s
19.9
tokens/s
6.17
sec
pp2048+tg256
716
tokens/s
21.8
tokens/s
2.90
sec
pp2048+tg768
714
tokens/s
21.1
tokens/s
2.91
sec
pp1024+tg1024
700
tokens/s
21.3
tokens/s
1.51
sec
pp1280+tg3072
720
tokens/s
20.3
tokens/s
1.83
sec
pp384+tg1152
756
tokens/s
22.9
tokens/s
550
ms
pp64+tg1024
557
tokens/s
23.7
tokens/s
155
ms
pp16+tg1536
227
tokens/s
23.4
tokens/s
113
ms