Stop Reasons¶

When generating a response, it's useful to know why the model stopped producing tokens.

Every completed message in AiKit includes a stopReason property that indicates how the generation ended:

• "stop" - Normal completion, the model finished its response.
• "length" - The output hit the maximum token limit.
• "toolUse" - The model is calling tools and expects tool results to be returned.
• "error" - An error occurred during generation.
• "cancel" - The request was cancelled.

const message = await stream.complete(model, context);

if (message.stopReason === "length") {
    console.warn("The model hit the token limit before finishing.");
} else if (message.stopReason === "tool-calls") {
    console.log("The model invoked tools.");
} else {
    console.log("Finished normally.");
}