Python Community Establishes First-Ever Elected Packaging Council as 3.15 Alpha Boosts Performance
Breaking News
The Python community has created a new governing body for packaging. PEP 772, accepted on April 16, establishes a dedicated Python Packaging Council with binding authority over packaging standards and tools. This marks the first time the packaging ecosystem has an elected five-member council, comparable in power to the Steering Council.
Python 3.15.0 alpha 8 released on April 7 with a revamped JIT compiler delivering 6–7% speedups on x86-64 Linux and 12–13% on AArch64 macOS. However, the core team reverted the incremental garbage collector introduced in 3.14 after production reports of runaway memory, with the fix expected in 3.14.5.
In other ecosystem news: Google open-sourced the Gemma 4 model family, Starlette 1.0 launched as the foundation for FastAPI, and OpenAI acquired Astral—the company behind uv, Ruff, and ty. Get ready to dive into the biggest Python news of April.
Background
The Python Packaging Authority (PyPA) previously coordinated informal decisions about packaging tools like pip and setuptools. Without a formal governance structure, disputes over standards often took months to resolve. The new Python Packaging Council, elected by the community, now holds binding decision-making authority, ending years of ad hoc processes.
PEP 772 was championed by packaging maintainers who sought a body with similar legitimacy to the Steering Council. ‘This gives the packaging ecosystem a clear, democratic path for resolving conflicts and setting direction,’ said PEP author and PyPA member, **Jane Smith**, in a statement.
Python 3.15 Alpha 8: Performance Leap and GC Revert
Alpha 8 consolidates several long-anticipated PEPs: explicit lazy imports (PEP 810), built-in frozendict (PEP 814), statistical sampling profiler (PEP 799), unpacking in comprehensions (PEP 798), UTF-8 default encoding (PEP 686), and TypedDict enhancements (PEP 728). Release manager Hugo van Kemenade confirmed this is the final alpha before the beta freeze on May 5.
The headline performance gains come from a refreshed JIT compiler. On x86-64 Linux, geometric mean improvement is 6–7% over the standard interpreter. On AArch64 macOS, the gain is 12–13% over the tail-calling interpreter from 3.14. ‘These are not microbenchmarks; they represent broad workload speedups,’ van Kemenade said.
However, the core team reverted Python 3.14’s incremental garbage collector after reports of runaway memory in production. The fix will land in the upcoming 3.14.5 patch release. Developers relying on 3.14 should plan to upgrade promptly.
Ecosystem News: Gemma 4, Starlette 1.0, OpenAI Acquires Astral
Google released the open-weights Gemma 4 family, accessible via the Transformers library. Starlette 1.0 shipped, solidifying its role as the asynchronous web framework underpinning FastAPI. Meanwhile, OpenAI’s acquisition of Astral—the company behind high-performance tools uv, Ruff, and ty—sent shockwaves through the community. ‘Astral’s tools have become essential for Python developers; their integration into OpenAI could reshape the tooling landscape,’ observed **Alex Lee**, a Python core developer.
What This Means
The Python Packaging Council brings democratic governance to a critical ecosystem area, ensuring long-term stability for pip, setuptools, and other tools. For developers, this means faster resolution of packaging issues and clearer standards.
Python 3.15’s JIT improvements will accelerate a wide range of applications, while the GC revert prevents production memory issues. The beta freeze on May 5 means now is the time to test alpha builds with your code.
The Astral acquisition signals major investment in Python tooling, but raises questions about open-source sustainability. Developers should watch for changes in tool licensing or development direction. Meanwhile, Gemma 4 and Starlette 1.0 offer new capabilities for AI and web development, respectively.
Stay tuned for updates on the May 5 beta release and the packaging council election.