High output tokens in a structured output query

[jesuspascuall]

Hello.
I've implemented a query to gpt5 API using a structured output (no more than 20 fields of less than 3 words each). When inspecting the token usage of my queries, I am using 1000 input tokens and 4000 output tokens for each query. Why am I getting that high output tokens?

create_payload = {
    "model": "gpt-5",
    "instructions": system_prompt,
    "input": [
        {
            "role": "user",
            "content": [
                {
                    "type": "input_text",
                    "text": "Extract the required fields and follow the JSON Schema.",
                },
                {"type": "input_file", "file_id": pdf_file.id},
            ],
        }
    ],
    "text": {
        "format": {
            "type": "json_schema",
            "name": "Extraction",
            "schema": schema_body,
            "strict": True,
        }
    },
}