Accuracy Distribution
Success rate based on wordbank.
01
GPT-5.6 Luna (xhigh)
77%
02
GPT-5.4 Mini (high)
77%
03
GPT-5.6 Luna (medium)
76%
04
Gemini 3 Flash (dynamic)
75%
05
GPT-5.6 Luna (high)
74%
06
GPT-5 Mini (medium)
71%
07
Gemini 3.1 Flash Lite (medium)
62%
08
GPT-5.4 Mini (medium)
60%
09
Gemini 3.1 Flash Lite (minimal)
46%
10
GPT-5.4 Nano (medium)
44%
11
Claude Haiku 4.5
33%