JSON Mode Intermediate

JSON mode constrains the model to output valid JSON. The provider guarantees that the response will be parseable JSON, eliminating the most common failure mode of structured output: invalid syntax.

OpenAI JSON Mode

Python 
from openai import OpenAI
import json

client = OpenAI()

# Basic JSON mode
response = client.chat.completions.create(
    model="gpt-4o",
    response_format={"type": "json_object"},
    messages=[
        {"role": "system", "content": "Extract data as JSON."},
        {"role": "user", "content": "Alice is 30 and lives in NYC."}
    ]
)

data = json.loads(response.choices[0].message.content)
# {"name": "Alice", "age": 30, "city": "NYC"}
OpenAI Requirement: When using response_format: {"type": "json_object"}, you must include the word "JSON" in either the system or user message. Otherwise, the API returns an error.

OpenAI Structured Outputs (Schema-Enforced)

OpenAI also offers a stricter mode that enforces a specific JSON Schema:

Python 
response = client.chat.completions.create(
    model="gpt-4o",
    response_format={
        "type": "json_schema",
        "json_schema": {
            "name": "person",
            "strict": True,
            "schema": {
                "type": "object",
                "properties": {
                    "name": {"type": "string"},
                    "age": {"type": "integer"},
                    "city": {"type": "string"}
                },
                "required": ["name", "age", "city"],
                "additionalProperties": False
            }
        }
    },
    messages=[
        {"role": "user", "content": "Alice is 30 and lives in NYC."}
    ]
)
# Guaranteed to match the schema exactly

Anthropic JSON via Tool Use

Anthropic does not have a dedicated JSON mode, but you can use tool use to get structured JSON output:

Python 
import anthropic

client = anthropic.Anthropic()

response = client.messages.create(
    model="claude-sonnet-4-20250514",
    max_tokens=1024,
    tools=[{
        "name": "extract_person",
        "description": "Extract person information from text.",
        "input_schema": {
            "type": "object",
            "properties": {
                "name": {"type": "string"},
                "age": {"type": "integer"},
                "city": {"type": "string"}
            },
            "required": ["name", "age", "city"]
        }
    }],
    tool_choice={"type": "tool", "name": "extract_person"},
    messages=[
        {"role": "user", "content": "Alice is 30 and lives in NYC."}
    ]
)

# Extract the structured data from tool_use block
tool_block = next(b for b in response.content if b.type == "tool_use")
data = tool_block.input  # {"name": "Alice", "age": 30, "city": "NYC"}

Google Gemini JSON Mode

Python 
import google.generativeai as genai

model = genai.GenerativeModel("gemini-2.0-flash")

response = model.generate_content(
    "Extract: Alice is 30 and lives in NYC.",
    generation_config=genai.GenerationConfig(
        response_mime_type="application/json",
        response_schema={
            "type": "object",
            "properties": {
                "name": {"type": "string"},
                "age": {"type": "integer"},
                "city": {"type": "string"}
            }
        }
    )
)

Comparison Table

Feature OpenAI JSON Mode OpenAI Structured Anthropic (Tool) Google
Valid JSON guaranteed Yes Yes Yes Yes
Schema enforcement No Yes Partial Yes
Requires schema No Yes Yes Optional
Best Practice: When possible, use schema-enforced structured output rather than plain JSON mode. Plain JSON mode guarantees valid JSON but not the correct structure. Schema enforcement guarantees both.
← Introduction Pydantic Output →