TEST #2052 RESULTS
12/27/2025 - 8:42 AM
ACCELERATOR
MODEL
100
tokens/s
generation
284
ms
time to first token
5436
tokens/s
prompt
1243
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
RAM
119.7GB
OS
Linux
Kernel Release
6.14.0-1015-nvidia
Architecture
aarch64
Version
Cosmopolitan 3.9.7 MODE=aarch64; #15-Ubuntu SMP PREEMPT_DYNAMIC Tue Nov 25 18:02:16 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
9079
tokens/s
118
tokens/s
120
ms
pp4096+tg256
3806
tokens/s
85.4
tokens/s
1.09
sec
pp2048+tg256
5113
tokens/s
97.9
tokens/s
413
ms
pp2048+tg768
4745
tokens/s
95.6
tokens/s
442
ms
pp1024+tg1024
6230
tokens/s
97.4
tokens/s
175
ms
pp1280+tg3072
6285
tokens/s
92.8
tokens/s
213
ms
pp384+tg1152
8499
tokens/s
102
tokens/s
54
ms
pp64+tg1024
3917
tokens/s
109
tokens/s
25
ms
pp16+tg1536
1254
tokens/s
105
tokens/s
22
ms