Models & thinking levels

Explore model configurations from direct model calls across recorded thinking levels. Provider release dates and specifications are listed when verified.

Model	Execution	Thinking	Acceptable rate
HUMAIN M3	Direct	default	67.35%
HUMAIN M3	Direct	disabled	62.54%
HUMAIN M3	Direct	high	66.86%
HUMAIN M3	Direct	low	66.87%
HUMAIN M3	Direct	medium	66.73%
HUMAIN M3	Direct	minimal	66.92%
HUMAIN M3	Direct	xhigh	67.06%
MiniMax M3	Direct	default	66.12%
MiniMax M3	Direct	disabled	61.83%
MiniMax M3	Direct	high	66.20%
MiniMax M3	Direct	low	66.18%
MiniMax M3	Direct	medium	65.97%
MiniMax M3	Direct	minimal	65.77%
MiniMax M3	Direct	xhigh	65.79%

Historical results. Each row represents a distinct configuration; scores are not pooled. Historical evaluation and publication dates have not been verified.