TEST #3496 RESULTS
05/10/2026 - 8:13 PM
ACCELERATOR
MODEL
52.2
tokens/s
generation
430
ms
time to first token
2963
tokens/s
prompt
711
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Xeon CPU E5-1630 v3 @ 3.70GHz (haswell)
RAM
127.9GB
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
3682
tokens/s
45.1
tokens/s
303
ms
pp4096+tg256
2976
tokens/s
51.9
tokens/s
1.39
sec
pp2048+tg256
3433
tokens/s
56.5
tokens/s
612
ms
pp2048+tg768
3348
tokens/s
51.5
tokens/s
627
ms
pp1024+tg1024
3644
tokens/s
55.4
tokens/s
300
ms
pp1280+tg3072
3581
tokens/s
52.8
tokens/s
382
ms
pp384+tg1152
3566
tokens/s
52.1
tokens/s
130
ms
pp64+tg1024
2143
tokens/s
52.1
tokens/s
46
ms
pp16+tg1536
297
tokens/s
52.0
tokens/s
77
ms