TEST #1910 RESULTS
12/03/2025 - 9:21 PM
ACCELERATOR
MODEL
41.0
tokens/s
generation
513
ms
time to first token
2865
tokens/s
prompt
612
LocalScore
HOW YOU STACK UP
Explore All ResultsQwen3 4B Thinking 2507 - Q8_0
SYSTEM
CPU
Intel Core i9-14900HX (alderlake)
RAM
95.7GB
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
3698
tokens/s
41.2
tokens/s
301
ms
pp4096+tg256
2156
tokens/s
38.6
tokens/s
1.93
sec
pp2048+tg256
2985
tokens/s
40.7
tokens/s
710
ms
pp2048+tg768
2914
tokens/s
40.6
tokens/s
728
ms
pp1024+tg1024
3232
tokens/s
41.4
tokens/s
342
ms
pp1280+tg3072
3521
tokens/s
40.1
tokens/s
387
ms
pp384+tg1152
4840
tokens/s
42.2
tokens/s
105
ms
pp64+tg1024
1957
tokens/s
42.3
tokens/s
57
ms
pp16+tg1536
486
tokens/s
42.1
tokens/s
57
ms