OneJev1.6g-31B is a System One decision model built on google/gemma-4-31B-it. Give it a piece of text along with a few typed questions, and it returns a calibrated probability for every option in a single forward pass.

Quick start

uv venv -p 3.12 && source .venv/bin/activate
uv pip install flide torchvision
flide serve --model OmniJev/OneJev1.6g-31B

In another terminal:

curl http://localhost:8000/v1/systemone -H "Content-Type: application/json" -d @- <<EOF
{
  "state": {"ticket": "I was charged twice for order 4471 and the refund page shows an error."},
  "questions": {
    "refund": {"type": "noul", "instructions": "The customer wants money back"},
    "team": {"type": "choice", "instructions": "Which team should handle this ticket?",
             "criteria": {"billing": "billing", "tech": "technical support", "shipping": "shipping", "sales": "sales"}},
    "urgency": {"type": "score", "instructions": "How urgent is this ticket?",
                "criteria": ["low", "medium", "high", "critical"]}
  }
}
EOF

answers.refund.noul is the probability of yes, answers.team.probabilities gives one probability per option and answers.urgency.score is the expected level.

Flide (paper) serves TypeSafe's System One API. More examples are on GitHub.