Users report that long answers from an assistant sometimes stop in the middle of a sentence. Which explanation is most likely?

• AThe temperature is set too low, so the model runs out of plausible continuations.
• BThe context window is too small to hold the user's question.
• CThe maximum output length is capping generation before the model has finished.
• DThe model is hallucinating and abandoning the response.