Cpython 3.15 Jit Performance Deltas By Platform

Issue 76 Edition 2026-03-17 4 min read

Not accepted General

Sources: 1 • Confidence: Medium • Updated: 2026-03-18 14:28

Key takeaways

The CPython JIT met its stated (modest) performance goals more than a year early on macOS AArch64 and a few months early on x86_64 Linux.
Python 3.15’s JIT project status is described as back on track.
In Python 3.15 alpha on macOS AArch64, the JIT is approximately 11–12% faster than the tail-calling interpreter baseline.
In Python 3.15 alpha on x86_64 Linux, the JIT is approximately 5–6% faster than the standard interpreter baseline.

The CPython JIT met its stated (modest) performance goals more than a year early on macOS AArch64 and a few months early on x86_64 Linux.
In Python 3.15 alpha on macOS AArch64, the JIT is approximately 11–12% faster than the tail-calling interpreter baseline.
In Python 3.15 alpha on x86_64 Linux, the JIT is approximately 5–6% faster than the standard interpreter baseline.

What specific performance goals were used to declare the CPython JIT successful, and what metrics/benchmarks define those goals?
What benchmark suite, workload mix, and statistical methodology produced the 11–12% (macOS AArch64) and 5–6% (x86_64 Linux) speedup figures?
Do the reported speedups hold when including end-to-end costs such as warmup, compilation overhead, memory usage, and latency distribution (not just average runtime)?
What does 'back on track' concretely mean in terms of milestones (alphas, betas, RC), scope, and risk of regressions or rollbacks?
Which platforms/configurations are covered beyond macOS AArch64 and x86_64 Linux, and are there any materially different results elsewhere?

Earlier than expected CPython JIT performance progress may reduce perceived risk around Python 3.15 execution speed, potentially influencing timelines for ecosystem experimentation and adoption of the JIT in performance sensitive Python workloads.
Platform specific gains, larger on macOS AArch64 than x86_64 Linux, suggest uneven near term benefits by deployment environment, which could affect where performance focused Python usage concentrates first.

Publication of the exact performance goals, benchmarks, workload mix, and statistical method used to support the 11 to 12 percent and 5 to 6 percent speedup figures.
Results that include end to end costs such as warmup, compilation overhead, memory usage, and latency distribution, showing the speedups persist beyond average runtime.
Clear definition of back on track with concrete milestones for 3.15 alpha to beta to release candidate, and evidence of stable performance without regressions.

Revised or clarified benchmarks show materially smaller or inconsistent speedups across representative suites, or high variance that undermines the reported deltas.
End to end measurements show warmup or compilation overhead, memory growth, or tail latency offsets the runtime gains for typical workloads.
Project status changes from back on track to delays, scope reduction, or rollback due to regressions or platform issues beyond macOS AArch64 and x86_64 Linux.