TEST #2073 RESULTS
01/01/2026 - 8:54 PM
ACCELERATOR
MODEL
84.8
tokens/s
generation
229
ms
time to first token
5568
tokens/s
prompt
1272
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen2.5 14B Instruct - Q4_K - Medium
SYSTEM
CPU
AMD Ryzen Threadripper 3970X 32-Core Processor (znver2)
RAM
125.6GB
OS
Linux
Kernel Release
6.8.0-90-generic
Architecture
x86_64
Version
Cosmopolitan 3.9.7 MODE=x86_64; #91-Ubuntu SMP PREEMPT_DYNAMIC Tue Nov 18 14:14:30 UTC 2025
RUNTIME
Name
llamafile
Version
0.9.2
Commit Hash
a30b324
DETAILED RESULTS
TEST NAME
PROMPT
GENERATION
TTFT
pp1024+tg16
7212
tokens/s
85.0
tokens/s
154
ms
pp4096+tg256
5478
tokens/s
76.3
tokens/s
771
ms
pp2048+tg256
6574
tokens/s
82.2
tokens/s
324
ms
pp2048+tg768
6577
tokens/s
81.8
tokens/s
324
ms
pp1024+tg1024
7251
tokens/s
86.8
tokens/s
153
ms
pp1280+tg3072
6828
tokens/s
81.4
tokens/s
199
ms
pp384+tg1152
6860
tokens/s
89.1
tokens/s
67
ms
pp64+tg1024
2702
tokens/s
90.5
tokens/s
35
ms
pp16+tg1536
629
tokens/s
89.8
tokens/s
37
ms