TEST #1083 RESULTS
06/20/2025 - 12:40 AM
ACCELERATOR
MODEL
78.3
tokens/s
generation
403
ms
time to first token
3204
tokens/s
prompt
854
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama 3 8b Bnb 4bit - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 5800X3D 8-Core Processor (znver3)
RAM
63.9GB
OS
Windows
Kernel Release
10.0
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
3881
tokens/s
81.0
tokens/s
276
ms
pp4096+tg256
3057
tokens/s
77.0
tokens/s
1.35
sec
pp2048+tg256
3556
tokens/s
75.9
tokens/s
589
ms
pp2048+tg768
3560
tokens/s
76.1
tokens/s
590
ms
pp1024+tg1024
3872
tokens/s
79.3
tokens/s
277
ms
pp1280+tg3072
3751
tokens/s
77.3
tokens/s
357
ms
pp384+tg1152
3943
tokens/s
79.5
tokens/s
111
ms
pp64+tg1024
2512
tokens/s
78.7
tokens/s
38
ms
pp16+tg1536
700
tokens/s
79.6
tokens/s
36
ms