Python 3.15 Alpha 6 Launches With Major JIT Boost, UTF-8 Default, and New Profiler

Python 3.15 Alpha 6 Now Available—JIT Speeds Up by 8% on Mac

Python 3.15.0a6, the sixth of eight planned alpha releases, has been published by the Python core team. The update introduces a built-in statistical profiler, makes UTF-8 the default encoding, and delivers significant just-in-time (JIT) compiler optimizations.

“This alpha brings together months of work from dozens of contributors—especially the JIT team, who have achieved a 3–4% speedup on x86-64 Linux and 7–8% on AArch64 macOS,” said Hugo van Kemenade, release team member. “We encourage developers to test these features early and report any issues.”

Key Changes in This Alpha

• PEP 799 – A new high-frequency, low-overhead statistical sampling profiler and dedicated profiling package.
• PEP 798 – Unpacking in comprehensions now supports * and ** operators.
• PEP 686 – Python switches to UTF-8 as the default encoding for source files and I/O.
• PEP 782 – A new PyBytesWriter C API for constructing bytes objects more efficiently.
• PEP 728 – TypedDict gains support for typed extra items.
• JIT compiler overhaul – Geometric mean performance improvements of 3–4% on x86-64 Linux and 7–8% on AArch64 macOS compared to the standard interpreter.
• Improved error messages – Many tracebacks and syntax errors have been clarified.

Background: Why Alpha Releases Matter

Python 3.15 remains in active development. Alpha releases like 3.15.0a6 are intended for testing new features and bug fixes, not for production use. Features may be added until the beta phase begins on 2026-05-05, and modified or removed until the release candidate phase on 2026-07-28.

“The alpha window is the best time for the community to experiment with upcoming capabilities and give feedback,” explained van Kemenade. “We rely on reports from early testers to catch regressions before the beta cutoff.”

What This Means for Developers

For application developers, the adoption of UTF-8 as the default encoding will simplify cross‑platform code but may require adjustments for projects that relied on locale‑dependent behavior. The new profiler (PEP 799) offers a lightweight way to identify performance bottlenecks without heavy instrumentation.

Library authors should pay special attention to PEP 728 (TypedDict with extra items) and PEP 798 (comprehension unpacking). The JIT improvements are likely most beneficial for compute‑intensive workloads running on macOS ARM systems, where the observed speedup is largest.

What’s Next?

The next pre‑release, Python 3.15.0a7, is scheduled for 2026-03-10. The full 3.15 release schedule is detailed in PEP 790.

Developers can download this alpha from the official Python downloads page. Bugs should be reported on the CPython issue tracker.

“Thank you to all the volunteers who make this possible,” said van Kemenade. “Please consider supporting the Python Software Foundation through donations or organizational contributions.”