TEST #2797 RESULTS
02/23/2026 - 3:49 AM
ACCELERATOR
MODEL
45.3
tokens/s
generation
825
ms
time to first token
1806
tokens/s
prompt
463
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen3 4B Thinking 2507 - Q4_K - Medium
SYSTEM
CPU
Intel Core i5-14450HX (alderlake)
RAM
15.7GB
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
2080
tokens/s
49.0
tokens/s
513
ms
pp4096+tg256
1357
tokens/s
40.0
tokens/s
3.04
sec
pp2048+tg256
1739
tokens/s
45.2
tokens/s
1.20
sec
pp2048+tg768
1735
tokens/s
42.5
tokens/s
1.20
sec
pp1024+tg1024
2055
tokens/s
45.1
tokens/s
519
ms
pp1280+tg3072
1945
tokens/s
41.7
tokens/s
680
ms
pp384+tg1152
2630
tokens/s
46.6
tokens/s
167
ms
pp64+tg1024
2065
tokens/s
49.3
tokens/s
51
ms
pp16+tg1536
647
tokens/s
48.7
tokens/s
44
ms