TEST #2051 RESULTS
12/27/2025 - 8:39 AM
ACCELERATOR
MODEL
103
tokens/s
generation
305
ms
time to first token
5148
tokens/s
prompt
1202
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
6131
tokens/s
120
tokens/s
176
ms
pp4096+tg256
3321
tokens/s
88.0
tokens/s
1.25
sec
pp2048+tg256
4946
tokens/s
89.0
tokens/s
427
ms
pp2048+tg768
4793
tokens/s
98.3
tokens/s
438
ms
pp1024+tg1024
6065
tokens/s
99.8
tokens/s
178
ms
pp1280+tg3072
7630
tokens/s
96.7
tokens/s
177
ms
pp384+tg1152
8332
tokens/s
109
tokens/s
57
ms
pp64+tg1024
3936
tokens/s
114
tokens/s
26
ms
pp16+tg1536
1182
tokens/s
113
tokens/s
23
ms