TEST #2591 RESULTS
02/10/2026 - 5:12 PM
ACCELERATOR
MODEL
62.7
tokens/s
generation
600
ms
time to first token
2198
tokens/s
prompt
612
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
12th Gen Intel Core i7-12700KF (alderlake)
RAM
63.9GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.3
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2610
tokens/s
63.6
tokens/s
408
ms
pp4096+tg256
2002
tokens/s
58.5
tokens/s
2.06
sec
pp2048+tg256
2409
tokens/s
62.9
tokens/s
866
ms
pp2048+tg768
2363
tokens/s
62.2
tokens/s
883
ms
pp1024+tg1024
2611
tokens/s
63.4
tokens/s
408
ms
pp1280+tg3072
2532
tokens/s
61.1
tokens/s
521
ms
pp384+tg1152
2678
tokens/s
64.0
tokens/s
159
ms
pp64+tg1024
1965
tokens/s
64.2
tokens/s
48
ms
pp16+tg1536
609
tokens/s
64.6
tokens/s
42
ms