TEST #12 RESULTS
03/19/2025 - 10:54 PM
ACCELERATOR
MODEL
109
tokens/s
generation
443
ms
time to first token
3031
tokens/s
prompt
907
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD EPYC 7352 24-Core Processor (znver2)
RAM
125.64GB
OS
Linux
Kernel Release
6.8.0-52-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #53~22.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Jan 15 19:18:46 UTC 2
RUNTIME
Name
llamafile
Version
0.9.1
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3656
tokens/s
125
tokens/s
288
ms
pp4096+tg256
2623
tokens/s
64.4
tokens/s
1.58
sec
pp2048+tg256
3227
tokens/s
98.2
tokens/s
645
ms
pp2048+tg768
3231
tokens/s
93.6
tokens/s
644
ms
pp1024+tg1024
3559
tokens/s
115
tokens/s
296
ms
pp1280+tg3072
3479
tokens/s
84.1
tokens/s
376
ms
pp384+tg1152
3957
tokens/s
128
tokens/s
104
ms
pp64+tg1024
2311
tokens/s
141
tokens/s
34
ms
pp16+tg1536
1240
tokens/s
133
tokens/s
19
ms