TEST #2305 RESULTS
01/15/2026 - 6:35 PM
ACCELERATOR
36.5
tokens/s
generation
1.22
sec
time to first token
1393
tokens/s
prompt
347
LocalScore
HOW YOU STACK UP
Explore All ResultsLlama3.3 8B Instruct Thinking Heretic Uncensored Claude 4.5 Opus High Reasoning - Q6_K
SYSTEM
CPU
AMD Ryzen 5 5600 6-Core Processor (znver3)
RAM
31.9GB
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
239
tokens/s
12.4
tokens/s
4.36
sec
pp4096+tg256
1582
tokens/s
38.2
tokens/s
2.62
sec
pp2048+tg256
1800
tokens/s
39.5
tokens/s
1.17
sec
pp2048+tg768
1797
tokens/s
39.4
tokens/s
1.16
sec
pp1024+tg1024
1850
tokens/s
39.6
tokens/s
579
ms
pp1280+tg3072
1838
tokens/s
39.0
tokens/s
722
ms
pp384+tg1152
1959
tokens/s
40.0
tokens/s
223
ms
pp64+tg1024
1181
tokens/s
40.0
tokens/s
79
ms
pp16+tg1536
290
tokens/s
40.2
tokens/s
78
ms