Benchmarks & research

Janus

@FirasSX9142PythonMITupdated 2026-09-18runnable

Measure when to use Jev and other models on your data, then route accordingly.

FirasSX914/Janus

Where it calls Jev

from typesafe_sdk import TypeSafeClient  # import tardif : dependance optionnelle

src/janus/providers/typesafe.py:28

The link points at the commit we read, so the line number still holds.

What it asks Jev

These question sets are lifted from this project's source exactly as written.

from experiments/probe.py:26

  1. departmentchoice

    Quelle équipe doit traiter ce message ?

    • shippingSuivi de commande, retard, colis perdu
    • billingFacturation, paiement, remboursement

Run it on your own text

The code
import { experimental_evaluate as evaluate } from 'ai';

const { answers } = await evaluate({
  model: 'typesafe-ai/jev',
  state,
  questions: {
    department: { type: 'choice', instructions: 'Quelle équipe doit traiter ce message ?',
      criteria: { shipping: 'Suivi de commande, retard, colis perdu', billing: 'Facturation, paiement, remboursement' } },
  },
});

Other projects in this category