joey-1 decision model available on Baseten

joey-1 is now available through Baseten Model APIs, and its weights are open source on Hugging Face. It's a 2B-parameter decision model post-trained from Qwen3.5-2B that classifies text, routes requests, and scores content against a rubric with typed answers your application can use directly.

Include several questions in one request to get choices, yes/no probabilities, and scores. Questions about the same state reuse its input prefix, so they cost less together than in separate requests.

Call decision-models/baseten-joey-2b-latest through the System One endpoint with your Baseten API key:

1import os
2
3import requests
4
5resp = requests.post(
6    "https://inference.baseten.co/v1/systemone",
7    headers={"Authorization": f"Bearer {os.environ['BASETEN_API_KEY']}"},
8    json={
9        "model": "decision-models/baseten-joey-2b-latest",
10        "state": "I was charged twice for my subscription. Please refund the duplicate charge.",
11        "questions": {
12            "route": {
13                "type": "choice",
14                "instructions": "Which team should handle this request?",
15                "criteria": {
16                    "billing": "Billing, payments, and refunds",
17                    "technical_support": "Technical troubleshooting",
18                },
19            },
20            "refund": {
21                "type": "noul",
22                "instructions": "Is the customer asking for a refund?",
23            },
24            "urgency": {
25                "type": "score",
26                "instructions": "How urgent is the request?",
27                "criteria": [
28                    "Routine request",
29                    "Needs attention today",
30                    "Immediate emergency",
31                ],
32            },
33        },
34    },
35)
36
37print(resp.text)

For more information, see our docs.

joey-1 decision model available on Baseten

Explore Baseten today