TEST #665 RESULTS
04/10/2025 - 9:03 AM
ACCELERATOR
MODEL
24.9
tokens/s
generation
2.15
sec
time to first token
716
tokens/s
prompt
202
LocalScore
HOW YOU STACK UP
Explore All ResultsGemma 3 4b It - Q6_K
SYSTEM
CPU
AMD Ryzen 7 4800H with Radeon Graphics (znver2)
RAM
30.8GB
OS
Linux
Kernel Release
6.11.2-666.rog.fc42.x86_64
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Sun Oct 6 10:48:44 UTC 2024
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
852
tokens/s
30.3
tokens/s
1.23
sec
pp4096+tg256
502
tokens/s
16.7
tokens/s
8.22
sec
pp2048+tg256
671
tokens/s
21.0
tokens/s
3.10
sec
pp2048+tg768
660
tokens/s
19.5
tokens/s
3.15
sec
pp1024+tg1024
787
tokens/s
24.5
tokens/s
1.34
sec
pp1280+tg3072
745
tokens/s
17.9
tokens/s
1.76
sec
pp384+tg1152
933
tokens/s
29.3
tokens/s
438
ms
pp64+tg1024
862
tokens/s
34.0
tokens/s
98
ms
pp16+tg1536
433
tokens/s
31.1
tokens/s
60
ms