TEST #3989 RESULTS
08/03/2026 - 12:57 PM
ACCELERATOR
MODEL
34.1
tokens/s
generation
4.07
sec
time to first token
331
tokens/s
prompt
140
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3.2 1B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-2697 v3 @ 2.60GHz (haswell)
RAM
251.8GB
OS
Linux
Kernel Release
7.0.14-5-pve
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC PMX 7.0.14-5 (2026-07-14T12:32Z)
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
239
tokens/s
35.5
tokens/s
4.31
sec
pp4096+tg256
321
tokens/s
30.7
tokens/s
12.80
sec
pp2048+tg256
328
tokens/s
33.1
tokens/s
6.27
sec
pp2048+tg768
316
tokens/s
33.1
tokens/s
6.52
sec
pp1024+tg1024
465
tokens/s
34.7
tokens/s
2.23
sec
pp1280+tg3072
431
tokens/s
33.1
tokens/s
3.00
sec
pp384+tg1152
334
tokens/s
35.3
tokens/s
1.18
sec
pp64+tg1024
282
tokens/s
35.7
tokens/s
255
ms
pp16+tg1536
266
tokens/s
35.4
tokens/s
87
ms