Python's experimental JIT will be removed: what changes for those who depend on it

What happened and what to do

According to the LWN article, Python's experimental just-in-time compiler, added to CPython in 2024 for version 3.13, is on its way to being removed. In June, the Python steering council suspended JIT work on the main CPython branch until an accepted PEP addressed maintenance and compatibility concerns, set clear success metrics and provided a timeline. The maintainers created PEP 836 to meet that request, but withdrew it after months of discussion. For now, the project, once promising, appears to have come to an end.

In practice, a company that enabled the experimental JIT in production, or plans to, should treat the feature as not guaranteed in future releases. Useful steps include: inventorying services running Python 3.13 or later with the JIT enabled; measuring the real performance gain on representative workloads; and preparing a rollback plan to the standard interpreter in case the option is removed. It is also worth building a test matrix and a dashboard tracking latency and CPU use per service, so a runtime upgrade does not change performance unnoticed.

How the consultancy can help

Diagnosis of your exposure to Python's experimental JIT, with a service inventory, measurement of real gains and a risk assessment of removal. We then implement a benchmark harness, performance regression tests and a migration plan to the standard interpreter, with ongoing operation and monitoring.

Next step

If your team depends on Python performance and wants to assess the impact of this change, send a short description of your case, including the services involved and the version in use. We will reply with a scoped proposal for diagnosis and implementation.