TEST #3879 RESULTS
07/09/2026 - 12:52 AM
ACCELERATOR
MODEL
62.4
tokens/s
generation
683
ms
time to first token
1910
tokens/s
prompt
559
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
12th Gen Intel Core i9-12900H (alderlake)
RAM
31GB
OS
Linux
Kernel Release
7.0.0-27-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #27-Ubuntu SMP PREEMPT_DYNAMIC Thu Jun 18 19:13:49 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2237
tokens/s
66.0
tokens/s
473
ms
pp4096+tg256
1811
tokens/s
57.2
tokens/s
2.28
sec
pp2048+tg256
2071
tokens/s
61.4
tokens/s
1.00
sec
pp2048+tg768
2059
tokens/s
61.1
tokens/s
1.01
sec
pp1024+tg1024
2195
tokens/s
63.3
tokens/s
482
ms
pp1280+tg3072
2141
tokens/s
59.5
tokens/s
613
ms
pp384+tg1152
2274
tokens/s
64.4
tokens/s
184
ms
pp64+tg1024
2009
tokens/s
64.6
tokens/s
46
ms
pp16+tg1536
390
tokens/s
63.5
tokens/s
55
ms