Open Models Almanac
Methodsinference & servingreasoning control

specific method · filed under inference & serving

Full chain-of-thought output

The complete reasoning process is exposed in the output, rather than only used internally or summarized.

The sources on this page only mention it in passing. The line above is the extraction's one-line gloss. See reasoning control for the methods filed beside it.

Also called Full chain-of-thought.

sources
2
model
1
lab adopt it
1
strongest
core
↑