Aquesta és una versió anterior. La versió actual és la 2026.2.3. PyCharm 2026.2.3 · Data de publicació:
Fitxers d’instal·lació 2025.2.1
| Arquitectura | Tipus | Àmbit | Mida del fitxer | Baixa | SHA256 |
|---|---|---|---|---|---|
| 64 bits | EXE (NSIS) | Usuari actual | — | download.jetbrains.com | VirusTotal |
| 64 bits | EXE (NSIS) | Tots els usuaris | — | download.jetbrains.com | VirusTotal |
| ARM64 | EXE (NSIS) | Usuari actual | — | download.jetbrains.com | VirusTotal |
| ARM64 | EXE (NSIS) | Tots els usuaris | — | download.jetbrains.com | VirusTotal |
L’enllaç i l’SHA256 provenen del manifest comunitari del Gestor de paquets del Windows. Convé comparar el hash del fitxer baixat abans d’executar-lo.
Notes de la versió
Release highlights:
- PyCharm now autodetects Python interpreters installed with pyenv.
- When you use the Extract Method refactoring, PyCharm will automatically infer and insert parameter type hints.
- Instead of hard-coding a path, you can now use built-in macros (like $PROJECT_DIR$) to define the working directory in your Python test configurations.
- During a remote debug session, the IDE will open your local project files instead of a copy of remote sources.
- PyCharm no longer incorrectly flags code after a finally block as unreachable if the try statement contains a nested if.
- The original PYTHONPATH entries are now preserved when running configurations with a Docker Compose interpreter.
- PyCharm no longer reports a false warning when @classmethod is not the last decorator applied.
- The remote debug server can now be stopped reliably.
- The debugger no longer gets stuck on Collecting data... when a Mock local variable exists in a frame.
AI Playground improvements:
- You can now use models via OpenRouter in AI Playground by adding your own API key, giving you access to more providers.
- You can now easily create a new playground session with the same model and settings.
- Model settings have been reorganized for an improved selection and parameter editing experience.
- Redesigned model filtering makes it easier to choose specific models from each provider.
See the full release notes
