Open Models Almanac
Methodsinference & servingreasoning control

specific method · filed under inference & serving

Generate-verify-refine

A test-time procedure generates a candidate, verifies it, and refines it, with one report noting a maximum revision cap.

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 generate-verify-refine loop, Generate-verify-refine test-time scaling, generate–verify–refine methodology.

sources
2
models
2
labs adopt it
2
strongest
used

Documented in

Evidence

2 quotes from 2 sources, one card per model, each opening with its strongest.

How sources treat it
used 2

DeepSeek-V3.2

DeepSeekadopted1 source

In the IMO and CMO tasks, we employ a generate-verify-refine loop.

usedevaluation onlyin DeepSeek-V3.2DeepSeek
↑