TEST #2802 RESULTS
02/23/2026 - 4:23 AM
ACCELERATOR
MODEL
30.4
tokens/s
generation
14.70
sec
time to first token
95
tokens/s
prompt
58
LocalScore
HOW YOU STACK UP
Explore All Resultsh2o-danube3-500m-chat - IQ1_M - 1.75 bpw
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
72
tokens/s
13.6
tokens/s
14.47
sec
pp4096+tg256
83
tokens/s
30.9
tokens/s
49.61
sec
pp2048+tg256
96
tokens/s
31.9
tokens/s
21.39
sec
pp2048+tg768
100
tokens/s
33.0
tokens/s
20.65
sec
pp1024+tg1024
109
tokens/s
32.5
tokens/s
9.49
sec
pp1280+tg3072
106
tokens/s
32.7
tokens/s
12.19
sec
pp384+tg1152
110
tokens/s
32.4
tokens/s
3.58
sec
pp64+tg1024
98
tokens/s
33.7
tokens/s
691
ms
pp16+tg1536
77
tokens/s
33.0
tokens/s
240
ms