TEST #1318 RESULTS
08/05/2025 - 11:20 PM
ACCELERATOR
MODEL
39.1
tokens/s
generation
402
ms
time to first token
3159
tokens/s
prompt
675
LocalScore
HOW YOU STACK UP
Explore All ResultsGemma-3-12B-It - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 9900X 12-Core Processor
RAM
92GB
OS
Linux
Kernel Release
6.8.0-63-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #66-Ubuntu SMP PREEMPT_DYNAMIC Fri Jun 13 20:25:30 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3921
tokens/s
43.5
tokens/s
284
ms
pp4096+tg256
3294
tokens/s
30.9
tokens/s
1.28
sec
pp2048+tg256
3668
tokens/s
37.2
tokens/s
587
ms
pp2048+tg768
3647
tokens/s
36.0
tokens/s
588
ms
pp1024+tg1024
3965
tokens/s
39.9
tokens/s
282
ms
pp1280+tg3072
3873
tokens/s
34.4
tokens/s
355
ms
pp384+tg1152
3747
tokens/s
42.2
tokens/s
125
ms
pp64+tg1024
1810
tokens/s
44.7
tokens/s
56
ms
pp16+tg1536
507
tokens/s
43.5
tokens/s
57
ms