TEST #3063 RESULTS
03/12/2026 - 12:43 AM
ACCELERATOR
MODEL
52.4
tokens/s
generation
896
ms
time to first token
1478
tokens/s
prompt
442
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
Intel Core i7-14700 (alderlake)
RAM
93.9GB
OS
Linux
Kernel Release
6.12.73+deb13-amd64
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #1 SMP PREEMPT_DYNAMIC Debian 6.12.73-1 (2026-02-17)
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
1688
tokens/s
54.5
tokens/s
625
ms
pp4096+tg256
1370
tokens/s
46.0
tokens/s
3.01
sec
pp2048+tg256
1564
tokens/s
51.2
tokens/s
1.33
sec
pp2048+tg768
1560
tokens/s
50.4
tokens/s
1.33
sec
pp1024+tg1024
1660
tokens/s
53.1
tokens/s
635
ms
pp1280+tg3072
1659
tokens/s
49.5
tokens/s
790
ms
pp384+tg1152
1797
tokens/s
55.0
tokens/s
232
ms
pp64+tg1024
1471
tokens/s
56.3
tokens/s
61
ms
pp16+tg1536
535
tokens/s
55.7
tokens/s
48
ms