model_router

Asynchronously try each request, falling back when one with images is rejected.

Try each request in order, falling back when one with images is rejected.

Select one value from a fixed, unordered set of options.

from langchain_openai.decisions import Choice, OpenAIDecisions

decisions = OpenAIDecisions(model="gpt-6-luna")
response = decisions.invoke(
    {
        "input": "I was charged twice for my order.",
        "questions": {
            "department": Choice(
                instructions="Which department should handle this?",
                choices={
                    "billing": "Payments, invoices, and refunds.",
                    "technical": "Problems using the product.",
                    "other": "Requests outside these categories.",
                },
            ),
        },
    }
)
department = response.choices["department"]
if department.confidence >= 0.7:
    route_to(department.choice)

Complete input for one OpenAIDecisions invocation.

Keeping the input and questions together in the Runnable input ensures both participate in composition, batching, and tracing.

Typed answers and metadata returned from one Decisions request.

Access every answer through answers, or use predicates, choices, scores, and refusals for views filtered by answer type. Each mapping is keyed by the question names supplied in DecisionRequest.questions.

Ask typed questions about text or images with the OpenAI Decisions API.

A single request can combine Predicate probabilities, Choice selections, and Score ratings over shared input. The response preserves probabilities, confidence, and token usage so application code can decide whether to act, route, or request human review.

Sync and async OpenAI clients are created on first use of invoke or ainvoke respectively.

A model available to the router and the criterion for selecting it.

Routing answer stored in agent state under model_route.

A plain mapping rather than a ChoiceAnswer, so checkpointers can persist it without registering custom types.

Select an agent's model with an OpenAI Decisions Choice question.

Classifies the latest human message once before each agent run, stores the answer as a ModelRoute in agent state under model_route, and uses the selected route for every model call in the run. Keeping the complete answer makes probabilities and confidence available in state and traces.

The message text and base64 images are classified. Hosted image URLs, files, and audio are replaced with placeholders such as [image omitted]. If a request with images is rejected (HTTP 400 or 413), routing is retried once with every image replaced.

When no route is selected, because the model refused or the state has no human message, model_route is set to None and model calls use the agent's own model. API errors propagate and terminate the run.