Planned Enhancements and Performance Goals for Python 3.13

Python 3.13 is expected to be released around October 1, 2024, with developers focusing on significant performance improvements for the CPython reference implementation.

The release will introduce a Tier‑2 optimizer, enable subinterpreter usage directly from Python code, and enhance memory management to reduce interpreter overhead by at least 50%.

Overview

The 3.13 plan mirrors early 3.12 goals but benefits from completed foundational work, including low‑impact monitoring (PEP 669), a more robust bytecode compiler state, an active interpreter generator, completed register‑machine experiments, and a viable copy‑and‑patch machine‑code generator.

Three parallelizable work streams are planned:

Tier‑2 optimizer

Enabling subinterpreters from Python code (PEP 554)

Memory management improvements

Tier‑2 Optimizer

The Tier‑2 optimizer aims to make the second‑level interpreter functional, generate (initially low‑quality) superblocks, manage them, and integrate parallel workflows such as build‑time integration, tier‑2 code generation, de‑optimization support, superblock creation enhancements, specialization, partial assigners, and a copy‑and‑patch code generator.

Enabling Subinterpreters from Python

This work builds on the per‑interpreter GIL enhancements introduced in Python 3.12, allowing Python programmers to leverage better parallelism without writing C extensions; the effort follows the draft PEP 554.

Better Memory Management

Analysis shows a substantial portion of runtime is spent on memory management and cyclic garbage collection; the plan includes researching and experimenting with improved data structures to reduce allocations, decreasing time spent on periodic GC, and potentially implementing a new incremental GC.

Reference links: Phoronix article , GitHub README .