Matilda-Jev vs Jev

Matilda-Jev is Maincode's open decision model, published on Hugging Face as matilda-jev-v1 and matilda-jev-v1.5. Like Jev, it's a decision checkpoint, not a text generator: you give it state and a typed question — a choice, a yes/no (noul), or an ordered score — and it reads out the option, in one pass. It accepts text or JSON and optional images, with a 255-option readout. Here's the honest comparison with hosted Jev, and where each fits.

If you found Matilda-Jev while looking at Jev, the family resemblance is the point: it's an open checkpoint that does the Jev-shaped job — bounded, typed decisions instead of prose. The choice is the usual one for the category: an open model you pull and run yourself (Matilda-Jev) versus a hosted, calibrated endpoint you call (Jev).

What Matilda-Jev is

Matilda-Jev is Maincode's one-pass decision model, distributed as open weights on Hugging Face (matilda-jev-v1 and the later matilda-jev-v1.5). It scores the options supplied in a choice, noul (yes/no), or ordered score question, takes text or JSON state plus optional images, and exposes a 255-option readout. Maincode frames it explicitly as a decision checkpoint rather than a text-generation checkpoint — so, like Jev, it emits a structured answer over your options, not sentences you have to parse. Because it ships as weights, you run it yourself; detailed specs and licence live on the model cards, linked below.

How it lines up with Jev

Both the Matilda-Jev and Jev columns reflect each project's own published description; treat them as vendor-reported, and verify the model-card details before you commit.

When to pick which

• Reach for Matilda-Jev when you specifically want open weights to run yourself — for on-prem or offline use, to inspect or fine-tune the checkpoint, or to avoid a per-call bill. Its optional image input also matters if your decision depends on a picture, which text-only Jev doesn't take.
• Reach for Jev when you want a hosted, RLCD-calibrated model with pinned versions and honest confidence, no model hosting to run, and a free browser playground to try one decision before you wire anything up.
• Since both answer the same choice / noul / score shape, you can prototype on one and keep the other as a fallback with little rework.

A note on evidence: Matilda-Jev is a newer, smaller-profile release than Jev, Clef or the big-lab deciders, so there's less independent benchmarking of it so far. Don't take any single composite — from either side — as the answer. Run your own decision on your own data, which is cheap to do on both.

Make one real decision on Jev

The fastest way to judge any decision model is to run your own decision through it. Jev's playground does it in the browser with no signup, and a jv_live_ key points your code at the hosted endpoint — the same choice / noul / score shape Matilda-Jev expects:

// Hosted, calibrated decision — the same choice / noul / score shape
// an open decision checkpoint like Matilda-Jev expects.
const res = await fetch("https://jevtypesafeai.com/api/v1/decide", {
  method: "POST",
  headers: {
    "Content-Type": "application/json",
    Authorization: `Bearer ${process.env.JEV_KEY}`, // jv_live_...
  },
  body: JSON.stringify({
    state: { message: input },
    questions: {
      intent: { type: "choice", instructions: "What does the user want?",
                criteria: { buy: "ready to purchase", support: "needs help", browse: "just looking" } },
      priority: { type: "score", instructions: "How urgent is it?", criteria: ["low", "medium", "high"] },
    },
  }),
});
const { answers } = await res.json();
handle(answers.intent.choice, answers.priority.score);

• Hugging Face: Maincode/matilda-jev-v1.5 — Maincode's model card for the later checkpoint — the authoritative source for specs and licence
• Hugging Face: Maincode/matilda-jev-v1 — The original matilda-jev checkpoint

FAQ

See also: Jev alternatives · Clef vs Jev · Strands Decider vs Jev · What is a decision model · Playground