- Sources: primary, discussion
- Summary: LWN, in an account dated 2026-09-01, reports that the JIT shipped experimentally in Python 3.13 on an informational PEP, and that in June the steering council ruled no new JIT development may land in main beyond bug and security fixes until a standards-track PEP is accepted, stating the JIT code is to be removed from main if none is accepted within six months. PEP 836 was drafted in response and the discussion stalled on whether it should be split, with several core developers arguing the acceptance criteria and the implementation plan belong in separate documents and one warning the pressure could drive contributors away. The PEP targets at least a 20 percent pyperformance geometric mean improvement with free-threaded Python by the first 3.17 beta and proposes replacing the tracing frontend with a method-based one, while building with the JIT still pins an LLVM version, 19 for 3.14 and 21 for 3.15, both already end of life, which PEP 774 would remove.
- Why it matters: A JIT that already ships in released CPython is on a removal clock, so anyone planning around it depends on the PEP 836 outcome rather than on the code in main.
- Follow-up: Track whether PEP 836 is accepted before the six-month removal threshold, on which the council has published no outcome.
send feedback on this story