TEST #4238 RESULTS
10/01/2026 - 8:15 PM
ACCELERATOR
MODEL
53.2
tokens/s
generation
4.60
sec
time to first token
280
tokens/s
prompt
148
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2698 v4 @ 2.20GHz (broadwell)
RAM
251.6GB
OS
Linux
Kernel Release
7.0.0-111031-tuxedo
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #31~24.04.1tux1 SMP PREEMPT_DYNAMIC Tue Sep 1 19:30:32 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
251
tokens/s
56.3
tokens/s
4.09
sec
pp4096+tg256
274
tokens/s
46.7
tokens/s
14.97
sec
pp2048+tg256
336
tokens/s
53.9
tokens/s
6.11
sec
pp2048+tg768
337
tokens/s
49.0
tokens/s
6.09
sec
pp1024+tg1024
237
tokens/s
53.4
tokens/s
4.34
sec
pp1280+tg3072
307
tokens/s
48.2
tokens/s
4.19
sec
pp384+tg1152
304
tokens/s
56.0
tokens/s
1.28
sec
pp64+tg1024
251
tokens/s
59.1
tokens/s
271
ms
pp16+tg1536
221
tokens/s
56.5
tokens/s
89
ms