NEW YORK — What happened: Following successful experimental rollouts in earlier versions, Python core contributors are evaluating whether Python 3.15 can safely enable the free-threaded build mode (PEP 703) in standard binary distributions for major operating systems.
Why it matters: Python powers millions of web backends, data science pipelines, and automation scripts. The Global Interpreter Lock (GIL) has historically restricted multi-threaded CPU-bound parallelism, forcing developers to rely on multiprocessing or external C extensions.
Deep dive: Removing or bypassing the GIL allows multiple native threads to execute Python bytecode concurrently on multi-core processors. While this unlocks massive performance gains for concurrent servers and data processing, it requires third-party extension authors to audit their C-extensions for thread safety, leading to a transitional period for popular ecosystem packages like NumPy and Pandas.
Report check (claims vs what is verified vs still rumor): It is verified that core developers are working toward making free-threading production-ready. It is unverified whether the standard installer will flip the default switch in 3.15 or wait until 3.16.
Open questions: How quickly will the broader data science ecosystem update their binary wheels to support thread-safe execution without segmentation faults?