NO - NO INDIVIDUAL RIG RUNS IT
Jev has no published quantization that fits a rig one person can buy, so there's no local-hardware recommendation and no break-even to compute. The honest answer is the per-provider API cost above.
Decision model. Reads typed options from the logits, returns typed answers with probabilities in one pass. Three shapes: yes/no probability, choice from a set, score on a scale. No text generation, nothing to parse. RLCD training against proper scoring rules.
See the
model card
for the full architecture notes and any cloud subscription plans.