TensorFlowEnd of Life (EOL) Dates & Support Status

Active

FrameworksLatest Version: 2.21.0Latest Release: Mar 6, 2026

TensorFlow is Google's open-source machine learning platform. Only the latest minor receives patches; older minors are unsupported without fixed EOL dates. TensorFlow 2.21 supports Python 3.10-3.13 and GPU builds use CUDA 12.5 with cuDNN 9.3; 2.14 used CUDA 11.8.

Support summary

TensorFlow currently has supported cycles. Track the next EOL date and validate upgrades before your change window.

What do EOL and support statuses mean?

Next EOL

No announced EOL date is currently listed.

Supported cycles

0 lts

Versions

Latest Version

2.21.0

Active Support

EOL

Lifecycle Timeline

Mar 22, 2023As of Sep 27, 2026

2.21Active · Date not announced

2.20EOL · Aug 13, 2026

2.19EOL · Ended; date not announced

2.18EOL · Ended; date not announced

2.17EOL · Ended; date not announced

2.16EOL · Ended; date not announced

2.15EOL · Ended; date not announced

2.14EOL · Ended; date not announced

2.13EOL · Ended; date not announced

2.12EOL · Ended; date not announced

ActiveSecurityEOLHistorical phase unknown

Bars show history up to the snapshot date. Gray intervals lack dated support boundaries; arrows indicate an unannounced EOL date, not a support guarantee.

Versions

Release	Release Date	Active Support	EOL	Latest Version	LTS	Status
2.21	Mar 6, 2026	Maintained	Date not announced	2.21.0	No	Active
2.20	Aug 13, 2025	Ended	Aug 13, 2026	2.20.0	No	EOL
2.19	Mar 12, 2025	Ended	Ended; date not announced	2.19.1	No	EOL
2.18	Oct 25, 2024	Ended	Ended; date not announced	2.18.1	No	EOL
2.17	Jul 11, 2024	Ended	Ended; date not announced	2.17.1	No	EOL
2.16	Mar 8, 2024	Ended	Ended; date not announced	2.16.2	No	EOL
2.15	Nov 13, 2023	Ended	Ended; date not announced	2.15.1	No	EOL
2.14	Sep 26, 2023	Ended	Ended; date not announced	2.14.1	No	EOL
2.13	Jul 6, 2023	Ended	Ended; date not announced	2.13.1	No	EOL
2.12	Mar 22, 2023	Ended	Ended; date not announced	2.12.1	No	EOL

Data sources

Sources

Official project releases

Last verified

Jun 24, 2026

Community-maintained lifecycle data. Verify against official release notes before production migration.

Embed this badge

Show TensorFlow's live support status in your README, docs, or dashboard. The badge is a static SVG served from eol.wiki and follows the lifecycle data above.

Markdown

[![TensorFlow end-of-life status](https://eol.wiki/badge/tensorflow.svg)](https://eol.wiki/tensorflow/)

HTML

<a href="https://eol.wiki/tensorflow/"><img src="https://eol.wiki/badge/tensorflow.svg" alt="TensorFlow end-of-life status" height="20"></a>

All badges & lifecycle API →

Frequently Asked Questions

Which versions of TensorFlow have reached end of life?

The following TensorFlow versions have reached end of life and no longer receive security updates: 2.20, 2.19, 2.18, 2.17, 2.16, 2.15, 2.14, 2.13 (and 1 more). It is recommended to upgrade to a supported version.

Is it safe to continue using TensorFlow 2.20?

TensorFlow 2.20 has reached end of life and no longer receives security patches or bug fixes. Continuing to use it may expose your systems to known vulnerabilities. We strongly recommend upgrading to a supported version.

What is the latest version of TensorFlow?

The latest version of TensorFlow is 2.21.0, released in the 2.21 release cycle.

How many versions of TensorFlow are currently supported?

TensorFlow currently has 1 actively supported version(s): 2.21.

When does TensorFlow 2.21 reach end of life?

TensorFlow 2.21 does not have a specific end-of-life date announced yet.

What should I do when TensorFlow reaches end of life?

When a TensorFlow version reaches end of life, you should: 1) Plan your migration to a supported version as soon as possible. 2) Review the release notes for breaking changes. 3) Test your applications thoroughly in a staging environment. 4) Update your dependencies to ensure compatibility with the new version.