TEST #3203 RESULTS
03/28/2026 - 2:19 PM
ACCELERATOR
MODEL
49.1
tokens/s
generation
356
ms
time to first token
3661
tokens/s
prompt
796
LocalScore
HOW YOU STACK UP
Explore All ResultsMeta Llama 3.1 8B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen 7 7800X3D 8-Core Processor (znver4)
RAM
61.9GB
OS
Linux
Kernel Release
6.18.7-76061807-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #202601231045~1769703228~24.04~cb87b5b SMP PREEMPT_DYNAMIC Thu J
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
4734
tokens/s
54.2
tokens/s
235
ms
pp4096+tg256
3484
tokens/s
40.2
tokens/s
1.20
sec
pp2048+tg256
4162
tokens/s
47.6
tokens/s
513
ms
pp2048+tg768
4167
tokens/s
46.3
tokens/s
513
ms
pp1024+tg1024
4742
tokens/s
49.9
tokens/s
235
ms
pp1280+tg3072
4457
tokens/s
44.3
tokens/s
307
ms
pp384+tg1152
4816
tokens/s
52.3
tokens/s
99
ms
pp64+tg1024
1828
tokens/s
53.8
tokens/s
52
ms
pp16+tg1536
558
tokens/s
52.9
tokens/s
47
ms