New SAV Run

Configure and launch an SAV V4.0 evaluation

Deployment Type

Choose where the inference happens.

Models for Triangulation

Select local models running on your machine. Best for privacy and cost zero.

Agents per Sub-Persona

We generate synthetic profiles on the fly. Tip: At least 1 agent per sub-persona ensures coverage across all 20 base archetypes. Total agents: 20 (1 × 20 Base Sub-Personas).

Persona Generation Engine

Tip: For higher variety, use a creative model like Gemini or Claude. For speed and cost zero, use a local model.

Participating Archetypes (0)
No personas defined! Create one first.

0 API calls estimated (≈ $0.00 blended cost)

Execution Volume

0 runs × 20 agents × 1 models

Payload Assumption

Calculated using historical average per model

Generation Runs to Evaluate0 / 0 available
Loading generation runs…