How Microsoft’s Faster CPython Team Is Boosting Python Performance by Up to 60%

Because of the rise of machine learning, Python is now the world’s most popular programming language, yet its biggest weakness is performance, which often requires high‑end hardware and heavy reliance on GPUs.

Python’s creator, Guido van Rossum, wanted a language more flexible than others and aimed for it to be twice as fast as its current performance. In 2020, CPython core developer Mark Shannon drafted a plan to increase CPython’s speed by roughly five times, dividing the effort into four stages, each targeting about a 50% speed boost. CPython serves as the reference implementation for other variants such as the data‑science‑focused Anaconda.

Microsoft recently published a blog titled “A Team at Microsoft is Helping Make Python Faster,” which details the work of its Faster CPython team.

The article notes that Python’s ranking has risen in indices like TIOBE and IEEE, but its versatility often draws criticism for performance. Performance‑sensitive software is frequently written in languages like C, prompting ongoing efforts to improve Python’s CPU performance.

In 2020, Mark Shannon’s implementation plan aimed to accelerate CPython by about fivefold, split into four phases each delivering roughly a 50% speed increase. Guido van Rossum remarked that such a task is daunting for a volunteer and suggested that Microsoft might hire Mark and a small team to support him.

Consequently, Microsoft formed a six‑engineer team that includes Guido, who returned from retirement in November 2020 as a distinguished engineer in the Developer Division. Guido emphasized that a diverse team with varied backgrounds makes better decisions and remains open to unexpected opportunities.

The team has already helped other core developers follow the plan, and Python 3.11 has achieved performance gains of 10‑60% in certain areas.

Team members share a consistent commitment to CPython work, meeting every Monday to discuss progress, which fosters strong collaboration and mutual understanding of each other’s strengths and weaknesses.

Having a shared vision makes it easier to plan for the future. Most of the features in 3.11 were proposed before the team formed; the team now focuses on integrating many performance enhancements into this release while also researching features for future versions. Shannon’s original four‑stage plan continues to evolve and is expected to be refined over the coming years.

The team’s development aligns with Microsoft’s long‑term commitment to the Python community, aiming to “add value” for users. Team members stress that this focus on community needs benefits everyone, and they appreciate the transparency of their work.

Reddit users have commented that Microsoft achieved something Google could not, highlighting the significance of the effort.