TEST #1183 RESULTS
07/06/2025 - 4:07 PM
ACCELERATOR
MODEL
45.8
tokens/s
generation
282
ms
time to first token
4682
tokens/s
prompt
912
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 9 9950X3D 16-Core Processor
RAM
62.3GB
OS
Linux
Kernel Release
6.11.0-29-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #29~24.04.1-Ubuntu SMP PREEMPT_DYNAMIC Thu Jun 26 14:16:59 UTC 2
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
6427
tokens/s
52.0
tokens/s
178
ms
pp4096+tg256
4344
tokens/s
29.8
tokens/s
978
ms
pp2048+tg256
5618
tokens/s
44.2
tokens/s
387
ms
pp2048+tg768
5598
tokens/s
40.8
tokens/s
389
ms
pp1024+tg1024
6405
tokens/s
46.7
tokens/s
180
ms
pp1280+tg3072
5694
tokens/s
37.6
tokens/s
245
ms
pp384+tg1152
5871
tokens/s
51.5
tokens/s
83
ms
pp64+tg1024
1642
tokens/s
55.7
tokens/s
55
ms
pp16+tg1536
541
tokens/s
53.4
tokens/s
45
ms