Concepts
The ideas underneath the API
Jev launched with big claims, real evidence, and a gap between the two that is worth understanding precisely. These pages separate what is structural from what is measured, and what is measured from what is asserted.
What a System One model is
A new model class, or a classifier with better marketing? The honest answer is somewhere in between, and the distinction that matters is narrower than either camp says.
ReadCalibration, honestly
The central claim of the whole product, stated precisely — and what independent testers found when they finally measured it.
ReadWhat "cannot hallucinate" actually guarantees
A real, structural, useful guarantee — about the type of the answer, and nothing whatsoever about its correctness.
ReadRLCD, and what is actually known about it
A third post-training branch alongside RLHF and RLVR — announced with a name, an objective, and no paper.
ReadWhen not to use Jev
Fine-tuned classifiers, JSON mode, constrained decoding and embeddings all still exist. Here is the honest comparison.
Read