●Change8

Migrating to Pydantic v2.11.0b1

Version v2.11.0b1 introduces 5 breaking changes. This guide details how to update your code.

Released: 3/6/2025

5
Breaking Changes
5
Migration Steps
9
Affected Symbols

⚠️ Check Your Code

If you use any of these symbols, you need to read this guide:

GenerateSchemaFieldInfoTypeAdapterModelMetaclassClickHouseDsncheck_pydantic_core_versionBaseModel.model_fieldsBaseModel.model_computed_fieldsGenerateJsonSchema.literal_schema

Breaking Changes

●Issue #1

Fields annotated with `typing.Final` that also have a default value are now treated as private model attributes instead of class variables.

✓Solution

If you intend for the attribute to remain a class variable (not part of the model's data/schema), use `typing.ClassVar` instead of `typing.Final` for the annotation.

●Issue #2

Accessing `model_fields` or `model_computed_fields` directly on an instance of a Pydantic model is deprecated.

✓Solution

Access these attributes on the model class itself (e.g., `MyModel.model_fields`) instead of the instance (e.g., `instance.model_fields`).

●Issue #3

Core schema generation logic for path types has been moved inside the `GenerateSchema` class.

✓Solution

This is an internal change. If you were relying on internal Pydantic schema generation APIs related to path types, you will need to update your usage to reflect the new location within `GenerateSchema`.

●Issue #4

Support for Python 3.8 has been removed.

✓Solution

Ensure your project targets Python 3.9 or newer.

●Issue #5

The format for defining fields when using `create_model` has been reworked.

✓Solution

Review your usage of `pydantic.create_model` and update how field definitions (especially those involving defaults or metadata) are passed.

Migration Steps

  1. 1
    Review all model fields annotated with `typing.Final` that possess a default value. If these should remain class variables, change the annotation from `Final[T] = value` to `ClassVar[T] = value`.
  2. 2
    Update any code that accesses `model_fields` or `model_computed_fields` on model instances to access these attributes on the model class instead (e.g., change `instance.model_fields` to `ModelName.model_fields`).
  3. 3
    If your project relies on Python 3.8, upgrade your environment to Python 3.9 or later, as 3.8 support has been dropped.
  4. 4
    If you use `pydantic.create_model`, check the documentation for v2.11.0b1 regarding the reworked field definition format and adjust your calls accordingly.
  5. 5
    If you encounter issues related to schema generation for path types, consult the Pydantic source or documentation regarding the internal `GenerateSchema` class.

Release Summary

Pydantic v2.11.0b1 drops Python 3.8 support, introduces full PEP 695 support, and optimizes performance through improved caching and schema building. It also refactors internal schema generation and updates the behavior of Final attributes.

Need More Details?

View the full release notes and all changes for Pydantic v2.11.0b1.

View Full Changelog