TEST #2760 RESULTS
02/19/2026 - 5:03 PM
ACCELERATOR
MODEL
32.0
tokens/s
generation
6.43
sec
time to first token
221
tokens/s
prompt
103
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 5800X 8-Core Processor (znver3)
RAM
47.9GB
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
236
tokens/s
34.8
tokens/s
4.37
sec
pp4096+tg256
185
tokens/s
27.7
tokens/s
22.21
sec
pp2048+tg256
217
tokens/s
31.8
tokens/s
9.45
sec
pp2048+tg768
214
tokens/s
30.3
tokens/s
9.59
sec
pp1024+tg1024
225
tokens/s
33.3
tokens/s
4.58
sec
pp1280+tg3072
229
tokens/s
29.9
tokens/s
5.62
sec
pp384+tg1152
240
tokens/s
31.9
tokens/s
1.62
sec
pp64+tg1024
239
tokens/s
35.7
tokens/s
294
ms
pp16+tg1536
206
tokens/s
32.8
tokens/s
107
ms