Python releases version 3.14 - with cautious free-threaded support

The Python team has released version 3.14, with big new features including free threading support, the ability to use concurrent interpreters, improved debugger support, and an opt-in new interpreter which improves performance by 3 to 5 percent.

Free threading in Python, which disables the global interpreter lock (GIL), is now a complete implementation of PEP (Python Enhancement Proposal) 703, a much anticipated feature which makes concurrent programming in Python natural. Free-threaded mode also enables a specialized adaptive interpreter, originally part of the Faster CPython project led by Mark Shannon at Microsoft (though the company axed its support in May).

That said, the approach to free-threaded Python remains cautious. The big problem is that existing code may break (particularly when using the C API since it is the application binary interface, or ABI, that is not compatible), and a second issue is that single-threaded code runs between 5 to 10 percent slower when using the free-threaded interpreter.

The consequence is that although free-threaded Python is now fully supported, the free-threaded interpreter is not installed by default. On the Mac, the installer requires it to be selected as a customized install, and on Windows, using a new preview Python install manager from the Windows Store, the user needs to add the free-threaded install using: py install 3.14t . Once installed, the free-threaded build must be specified with a command such as python3.14t, otherwise the single-threaded build will run.

Other goodies in Python 3.14 include:

While the JIT compiler sounds exciting, it is not recommended for production use and can actually slow down performance is some cases, as well as not working well with some debuggers. Free-threaded Python builds do not support the JIT compiler.

Full details of what is new in Python 3.14 are in the release notes.®