TEST #977 RESULTS
06/03/2025 - 2:24 AM
ACCELERATOR
MODEL
37.0
tokens/s
generation
865
ms
time to first token
1421
tokens/s
prompt
393
LocalScore
HOW YOU STACK UP
Explore All ResultsGemma 3 4b It - Q4_K - Medium
SYSTEM
CPU
Intel Core i5-4590 CPU @ 3.30GHz (haswell)
RAM
31.9GB
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
1790
tokens/s
39.7
tokens/s
597
ms
pp4096+tg256
1539
tokens/s
34.0
tokens/s
2.69
sec
pp2048+tg256
1714
tokens/s
36.8
tokens/s
1.23
sec
pp2048+tg768
1707
tokens/s
36.3
tokens/s
1.23
sec
pp1024+tg1024
1795
tokens/s
37.4
tokens/s
595
ms
pp1280+tg3072
1751
tokens/s
35.2
tokens/s
755
ms
pp384+tg1152
1848
tokens/s
37.5
tokens/s
235
ms
pp64+tg1024
589
tokens/s
37.8
tokens/s
139
ms
pp16+tg1536
55
tokens/s
37.9
tokens/s
324
ms