Ladda ner PyCharm (EAP) 262.5752.27

Here are the highlights of the 2026.2 EAP 2 build that focuses on refining type inference for complex Python structures, improving the reliability of the debugpy engine, and stabilizing Jupyter workflows:
Code insight
- Improved Pydantic support: Field specifiers like Field() are now correctly recognized when used inside Annotated types (PY-89012). Added support for frozen Pydantic fields (PY-89180) and for validate_by_name and validate_by_alias config options (PY-89182).
- Decorator support: Improved typing stability by resolving a false compilation error that triggered when handling a ParamSpec type list in bounds - PY-89092.
- Refined generic unification: Fixed a regression where an empty list literal could cause generic unification to collapse to Any - PY-89047.
- Enhanced type checking for variadic arguments: PyCharm now performs individual type checking for arguments mapped to *args: *tuple[T1, T2, ...] - PY-88727.
- Callable type rendering: Type parameters are now correctly rendered for Callable types when using generics - PY-88621.
Debugger
- Reliable type rendering: Support for custom type renderers for debugpy was introduced, ensuring variables are displayed according to user-defined rules - PY-86931.
- Secure TLS verification: Resolved an issue where debugpy would fail with custom CA-signed TLS certificates - PY-89519.
Jupyter
- Stability in Colab: Fixed an issue where the Jupyter DAP debugger would fail to restart properly when working with Google Colab - PY-89463.
- Reliable cell rendering: Resolved a bug where markdown cells would occasionally become blank after being left idle - PY-89532.
Frameworks and tools
- Ruff integration: Fixed an issue with Ruff integration failure in workspace projects - PY-88814.
- UV remote execution: Resolved a regression that prevented uv run from functioning correctly for remote targets - PY-89538.

Full release notes.