The article “How fast is Python 3.15?” provides a detailed performance analysis of the upcoming CPython release, focusing on benchmark‑driven comparisons across standard CPython, the new JIT compiler, and the free‑threaded (no‑GIL) build. Using two deterministic workloads—recursive Fibonacci and bubble sort—the author evaluates Python 3.15 in both single‑threaded and multi‑threaded execution modes, generating a clear performance profile relevant for developers, performance engineers, and technical SEO audiences researching Python speed improvements.
In single‑threaded Fibonacci, Python 3.15 shows incremental gains over Python 3.14, continuing the trend of small but consistent optimizations introduced after the major jump in Python 3.11. The standout result is the Python 3.15 JIT, which delivers a 1.20× speedup over standard CPython. This marks the first release where the JIT consistently outperforms the interpreter in this benchmark. PyPy remains significantly faster, achieving 5.5× the speed of CPython 3.15, while Rust demonstrates extreme performance at 77× faster, reinforcing the gap between dynamic and compiled languages.
The single‑threaded bubble‑sort benchmark confirms the same pattern: Python 3.15 is slightly faster than 3.14, but the JIT build provides a substantial 1.28× improvement. Node.js surpasses PyPy in this specific workload, highlighting how algorithmic characteristics influence interpreter performance. Free‑threaded CPython remains slower in single‑thread mode due to concurrency‑related overhead, which is expected.
The multi‑threaded benchmarks reveal the most important performance evolution. Standard CPython continues to scale poorly because of the Global Interpreter Lock (GIL), and Python 3.15 is marginally slower than 3.14 in multi‑threaded Fibonacci. PyPy also suffers from GIL‑related limitations.However, free‑threaded Python 3.15 demonstrates a major breakthrough: in multi‑threaded Fibonacci, it achieves a 4.5× speedup compared to standard CPython, matching and stabilizing the gains introduced in Python 3.14 FT. This confirms that the no‑GIL architecture is maturing and becoming a viable option for CPU‑bound parallel workloads. The JIT build maintains its advantage even under multi‑threading, showing consistent performance improvements across all categories.
From a technical SEO perspective, the article positions Python 3.15 as a release focused on predictable single‑threaded optimizations, significant JIT compiler progress, and high‑impact concurrency improvements through the free‑threaded runtime. These advancements make Python 3.15 a strategically important version for developers targeting performance‑critical applications, parallel computing, and future no‑GIL migration paths.
See the full article on the official blog.