TEST #1867 RESULTS
11/26/2025 - 6:42 AM
ACCELERATOR
MODEL
55.6
tokens/s
generation
963
ms
time to first token
1316
tokens/s
prompt
423
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon E-2144G CPU @ 3.60GHz (skylake)
RAM
62.7GB
OS
Linux
Kernel Release
6.14.0-36-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #36~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Wed Oct 15 15:45:17 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1554
tokens/s
59.0
tokens/s
675
ms
pp4096+tg256
1299
tokens/s
46.8
tokens/s
3.17
sec
pp2048+tg256
1455
tokens/s
54.0
tokens/s
1.43
sec
pp2048+tg768
1453
tokens/s
52.9
tokens/s
1.43
sec
pp1024+tg1024
1539
tokens/s
56.3
tokens/s
682
ms
pp1280+tg3072
1499
tokens/s
51.3
tokens/s
871
ms
pp384+tg1152
1547
tokens/s
59.0
tokens/s
264
ms
pp64+tg1024
1270
tokens/s
60.8
tokens/s
66
ms
pp16+tg1536
232
tokens/s
59.8
tokens/s
84
ms