TEST #1639 RESULTS
10/12/2025 - 10:53 PM
ACCELERATOR
MODEL
105
tokens/s
generation
75
ms
time to first token
17167
tokens/s
prompt
2880
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 9354 32-Core Processor (znver4)
RAM
1133.5GB
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
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
24149
tokens/s
110
tokens/s
51
ms
pp4096+tg256
19827
tokens/s
96.7
tokens/s
217
ms
pp2048+tg256
21911
tokens/s
99.0
tokens/s
104
ms
pp2048+tg768
21864
tokens/s
98.6
tokens/s
104
ms
pp1024+tg1024
24226
tokens/s
106
tokens/s
51
ms
pp1280+tg3072
21453
tokens/s
99.5
tokens/s
69
ms
pp384+tg1152
15770
tokens/s
111
tokens/s
37
ms
pp64+tg1024
4164
tokens/s
111
tokens/s
23
ms
pp16+tg1536
1142
tokens/s
111
tokens/s
22
ms