TEST #3252 RESULTS
04/07/2026 - 4:32 PM
ACCELERATOR
MODEL
48.6
tokens/s
generation
3.61
sec
time to first token
397
tokens/s
prompt
175
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 5800X3D 8-Core Processor (znver3)
RAM
62.7GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
450
tokens/s
52.8
tokens/s
2.29
sec
pp4096+tg256
324
tokens/s
42.4
tokens/s
12.64
sec
pp2048+tg256
398
tokens/s
48.3
tokens/s
5.17
sec
pp2048+tg768
359
tokens/s
46.0
tokens/s
5.74
sec
pp1024+tg1024
431
tokens/s
48.1
tokens/s
2.40
sec
pp1280+tg3072
414
tokens/s
44.9
tokens/s
3.11
sec
pp384+tg1152
478
tokens/s
51.3
tokens/s
819
ms
pp64+tg1024
389
tokens/s
52.2
tokens/s
201
ms
pp16+tg1536
332
tokens/s
51.7
tokens/s
64
ms