Content
Tag
1 entry tagged Reasoning · 1 term.
An inference behavior where a frontier model self-scales its reasoning depth and tool-use budget to match the complexity of the prompt, without the operator flipping a manual extended-thinking or test-time-compute switch.