<div> Virtual Private Networks (VPNs) have become essential tools for internet privacy, security, and circumventing geographic restrictions. However, systematic, long-term performance data measured from speci&#x1c;c geographic regions remains scarce in the academic literature. We present the Tokyo VPN Speed Monitor dataset, a comprehensive longitudinal collection of VPN performance metrics measured from Tokyo, Japan. The dataset encompasses 15 major commercial VPN services with automated measurements conducted every 6 hours over a period from December 9, 2025 to January 9, 2026 (approximately 32 days). Key metrics include download speed, upload speed, ping latency, stability scores calculated from 7-day rolling standard deviations, and daily pricing information. The data collection infrastructure comprises 8 automated engines built on Google Apps Script, ensuring consistent, reproducible, and maintenance-free measurements. All data and source code are released under the MIT License and are available through 7 DOI-assigned repositories including Harvard Dataverse, Zenodo, &#x1c;gshare, OSF, Mendeley Data, Kaggle, and IEEE DataPort. </div>