First Check

• I added a very descriptive title here.
• I used the GitHub search to find a similar question and didn't find it.
• I searched the FastAPI documentation, with the integrated search.
• I already searched in Google "How to X in FastAPI" and didn't find any information.
• I already read and followed all the tutorial in the docs and didn't find an answer.
• I already checked if it is not related to FastAPI but to Pydantic.
• I already checked if it is not related to FastAPI but to Swagger UI.
• I already checked if it is not related to FastAPI but to ReDoc.

Commit to Help

• I commit to help with one of those options 👆

Example Code

-

Description

FastAPI provides options to serialize Pydantic models to JSON using multiple different libraries, including stdlib json, ujson, orjson:
https://fastapi.tiangolo.com/advanced/custom-response/

But deserialization of HTTP request body from JSON format is done using json.loads:

At the same time, Pydantic 2.x introduced methods TypeAdapter.validate_json(str | bytes) which uses Rust-bases JSON parser under the hood. It has higher performance than json.loads, orjson.loads and ujson.loads:

import json
from timeit import timeit

import orjson
import ujson
from pydantic import BaseModel, TypeAdapter


class PydanticModel(BaseModel):
    a: int
    b: int

PydanticModelAdapter = TypeAdapter(PydanticModel)

iterations = 3_000_000
data = {"a": 1, "b": 2}
json_data = '{"a": 1, "b": 2}'
model = PydanticModel(a=1, b=2)

checks = [
    "PydanticModel(**data)",
    "PydanticModel.model_validate(data)",
    "PydanticModelAdapter.validate_python(data)",
    "PydanticModel(**json.loads(json_data))",
    "PydanticModel.model_validate(json.loads(json_data))",
    "PydanticModel(**orjson.loads(json_data))",
    "PydanticModel.model_validate(orjson.loads(json_data))",
    "PydanticModel(**ujson.loads(json_data))",
    "PydanticModel.model_validate(ujson.loads(json_data))",
    "PydanticModel.model_validate_json(json_data)",
    "PydanticModelAdapter.validate_json(json_data)",
]

for check in checks:
    print(f"{check} =", timeit(check, globals=globals(), number=iterations))

PydanticModel(**data) = 3.067713697997533
PydanticModel.model_validate(data) = 4.0357147680006165
PydanticModelAdapter.validate_python(data) = 3.345809325001028

PydanticModel(**json.loads(json_data)) = 9.452866644998721
PydanticModel.model_validate(json.loads(json_data)) = 10.779246789999888

PydanticModel(**orjson.loads(json_data)) = 4.756399215002602
PydanticModel.model_validate(orjson.loads(json_data)) = 5.459393262997764

PydanticModel(**ujson.loads(json_data)) = 6.935222607000469
PydanticModel.model_validate(ujson.loads(json_data)) = 7.8213590790001035

PydanticModel.model_validate_json(json_data) = 4.414445899001294
PydanticModelAdapter.validate_json(json_data) = 3.0091406040010042

Using TypeAdapter.model_validate_json() is 2 times faster than using json.loads and requires no additional dependencies.

Is is possible to change FastAPI to deserialize JSON by default using this method?

Operating System

Linux

Operating System Details

Linux 6.15.8-1-MANJARO #1 SMP PREEMPT_DYNAMIC Thu, 24 Jul 2025 08:21:53 +0000 x86_64 GNU/Linu

FastAPI Version

0.115.14

Pydantic Version

2.11.7

Python Version

3.12.2

Additional Context

orjson==3.11.1
ujson==5.10.0