TEST #3089 RESULTS
03/15/2026 - 1:47 PM
ACCELERATOR
MODEL
75.4
tokens/s
generation
388
ms
time to first token
3476
tokens/s
prompt
878
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 7800X3D 8-Core Processor (znver4)
RAM
61.9GB
OS
Linux
Kernel Release
6.18.7-76061807-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #202601231045~1769703228~24.04~cb87b5b SMP PREEMPT_DYNAMIC Thu J
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
4224
tokens/s
88.7
tokens/s
254
ms
pp4096+tg256
3047
tokens/s
47.6
tokens/s
1.37
sec
pp2048+tg256
3738
tokens/s
66.8
tokens/s
563
ms
pp2048+tg768
3730
tokens/s
63.6
tokens/s
564
ms
pp1024+tg1024
4192
tokens/s
76.4
tokens/s
256
ms
pp1280+tg3072
3950
tokens/s
58.7
tokens/s
336
ms
pp384+tg1152
4319
tokens/s
88.2
tokens/s
99
ms
pp64+tg1024
2910
tokens/s
97.6
tokens/s
31
ms
pp16+tg1536
1171
tokens/s
91.3
tokens/s
23
ms