TEST #238 RESULTS
04/04/2025 - 7:26 PM
ACCELERATOR
MODEL
52.8
tokens/s
generation
641
ms
time to first token
2039
tokens/s
prompt
552
LocalScore
HOW YOU STACK UP
Explore All ResultsDeepSeek R1 Distill Llama 8B - Q6_K
SYSTEM
CPU
AMD Ryzen 7 5800X 8-Core Processor (znver3)
RAM
47GB
OS
Linux
Kernel Release
6.8.0-56-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #58-Ubuntu SMP PREEMPT_DYNAMIC Fri Feb 14 15:33:28 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
2395
tokens/s
54.2
tokens/s
446
ms
pp4096+tg256
1932
tokens/s
49.2
tokens/s
2.14
sec
pp2048+tg256
2219
tokens/s
52.4
tokens/s
942
ms
pp2048+tg768
2212
tokens/s
51.8
tokens/s
945
ms
pp1024+tg1024
2390
tokens/s
53.3
tokens/s
447
ms
pp1280+tg3072
2315
tokens/s
51.1
tokens/s
572
ms
pp384+tg1152
2439
tokens/s
54.1
tokens/s
176
ms
pp64+tg1024
1918
tokens/s
54.7
tokens/s
51
ms
pp16+tg1536
529
tokens/s
54.3
tokens/s
48
ms