K MEANS AI | Parallax Studio

Evaluation runs selected modes sequentially so comparisons stay controlled and easy to review.

Evaluation

Controlled comparisonUse the same tenant, persona, and prompt across selected execution modes. Evaluation results are held in page memory and saved as run references only.

Virtual tenant Persona Consensus threshold Sequential round limit

Execution modes Sequential BridgeShared-context reasoning across rounds. One-shot synthesisParallel model pass synthesized into one answer. One-shot confirmationParallel model pass evaluated for agreement.

Evaluation prompt

Evaluation Results

Session memory