OpenJev (Qwen3.5 4B) now available on Unity Gateway

The OpenJev (Qwen3.5 4B) model has been integrated into Databricks Unity Gateway as a pay-per-token service, enabling queries for classification, scoring, and structured answers. ### Use cases Imagine you need to automatically evaluate multiple-choice questions or classify yes/no answers with associated probabilities. With this addition, you can:- Evaluate multiple-choice and scoring questions.- Obtain structured answers and associated probabilities.- Integrate the model using the TypeSafe System One API. ### How to use To query the model using the TypeSafe System One API, use the corresponding endpoint in your workspace: python import requests import json token = "your_access_token" url = "https://<your-workspace>.databricks.com/serving-endpoints/system.ai.openjev-qwen35-4b/invocations" headers = { "Authorization": f"Bearer {token}", "Content-Type": "application/json" } data = { "inputs": ["Evaluate the following multiple-choice question: ..."] } response = requests.post(url, headers=headers, data=json.dumps(data)) print(response.json()) What happens in this code: 1. Defines the URL for the model serving endpoint in your workspace. 2. Sets up authentication headers and content type. 3. Sends a POST request with the question to be evaluated. 4. Prints the structured response returned by the model. ### Limitations and requirements Customers are responsible for ensuring compliance with applicable Databricks model use terms. Model availability depends on the region configured in Unity Gateway.