AS of 8 December 2025, Cloudflare’s Python Workers are described as having exceptionally fast cold starts, thanks to dedicated memory snapshots and a uv-first workflow that integrates with uv tooling and the pywrangler pipeline. The update confirms that Python support now spans packages from Pyodide, including many that rely on dynamic libraries, with a focus on a Python-native developer experience and rapid deployment to Cloudflare’s edge network, which extends to 330 locations across 125 countries.
Benchmark data cited in the post shows mean cold starts for loading common packages like httpx, fastapi and pydantic at 1.027 seconds on Cloudflare Python Workers, versus 2.502 seconds for AWS Lambda and 3.069 seconds for Google Cloud Run, effectively making the Cloudflare option 2.4x faster than Lambda and 3x faster than Cloud Run in these tests.
Tools such as pywrangler enable developers to run dev and deploy commands that automate dependency installation and packaging, while a free tier offers 100,000 requests per day and 10ms CPU time per invocation. The article also highlights features such as type hints generation and support for workflows, durable objects and long-running WebSocket servers to broaden what a Python Worker can do on the edge.