TEST #618 RESULTS
04/08/2025 - 9:09 PM
ACCELERATOR
MODEL
22.3
tokens/s
generation
268
ms
time to first token
6976
tokens/s
prompt
834
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen Threadripper PRO 7995WX (znver4)
RAM
503.1GB
OS
Linux
Kernel Release
6.8.0-55-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #57-Ubuntu SMP PREEMPT_DYNAMIC Wed Feb 12 23:42:21 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
9921
tokens/s
22.0
tokens/s
149
ms
pp4096+tg256
4136
tokens/s
34.3
tokens/s
1.02
sec
pp2048+tg256
6775
tokens/s
12.7
tokens/s
384
ms
pp2048+tg768
6776
tokens/s
11.5
tokens/s
384
ms
pp1024+tg1024
9938
tokens/s
17.2
tokens/s
149
ms
pp1280+tg3072
7444
tokens/s
11.2
tokens/s
227
ms
pp384+tg1152
11520
tokens/s
25.2
tokens/s
57
ms
pp64+tg1024
4922
tokens/s
37.0
tokens/s
25
ms
pp16+tg1536
1353
tokens/s
29.6
tokens/s
22
ms