TEST #4111 RESULTS
09/01/2026 - 8:32 AM
ACCELERATOR
MODEL
68.4
tokens/s
generation
220
ms
time to first token
5980
tokens/s
prompt
1229
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen Threadripper 9960X 24-Cores
RAM
125.2GB
OS
Linux
Kernel Release
6.8.0-136-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #136-Ubuntu SMP PREEMPT_DYNAMIC Wed Jul 1 21:53:05 UTC 2026
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
8079
tokens/s
78.4
tokens/s
140
ms
pp4096+tg256
5618
tokens/s
42.4
tokens/s
753
ms
pp2048+tg256
6783
tokens/s
58.0
tokens/s
319
ms
pp2048+tg768
6871
tokens/s
56.2
tokens/s
315
ms
pp1024+tg1024
8097
tokens/s
68.8
tokens/s
140
ms
pp1280+tg3072
7173
tokens/s
54.3
tokens/s
192
ms
pp384+tg1152
7841
tokens/s
82.4
tokens/s
59
ms
pp64+tg1024
2495
tokens/s
89.8
tokens/s
36
ms
pp16+tg1536
862
tokens/s
85.0
tokens/s
29
ms