Algorithm Updates · 2026

X (Twitter) Algorithm Update Timeline 2026

Tracking X's timeline algorithm, Grok-powered ranking, ad system, verification, and creator monetization changes.

← All platform timelines

What is an X (Twitter) algorithm update?

An X algorithm update is a change to how X ranks posts in the For You and Following feeds. Since 2025, X has handed the ranking pipeline over to Grok, xAI's AI model. Grok reads every post (text, images, video), watches videos, and personalizes rankings in real time. In January 2026, X published the complete Grok-powered algorithm source code publicly on github.com/xai-org/x-algorithm, the first time the entire feed system has been visible in this form.

This page is the X algorithm update timeline for 2026. We track ranking signal changes, engagement-weight adjustments, ad-system updates, and creator-monetization changes. X's algorithm iterates quickly and openly, this timeline captures both the public source-code releases and the practical implications from head of product Nikita Bier's public updates.

2026 X updates at a glance

Confirmed 2026 X (Twitter) algorithm and ranking changes. Click any update to jump to full detail below.

Full timeline

July 2026MEDIUM IMPACT

Mutual-follower boost in replies

X adjusted its For You and reply ranking to surface more posts from mutuals, accounts you follow who also follow you back, especially inside reply threads. Head of Product Nikita Bier announced the change, saying the mutual signal had been missing and made friends less visible in reply threads. The goal is a more community-driven feed instead of a wall of strangers.

What this means for you

For practices using X for peer-to-peer or B2B: engagement from accounts that follow you back is now materially more visible. Practical effect: replying to peers and prompting them to reply back builds mutual visibility over time. For B2C healthcare/wellness accounts: this doesn't help much because followers rarely follow you back, so lean on the Following feed and community lists rather than fighting for reply visibility.

Read the source ↗
H1 2026HIGH IMPACT

Grok-powered ranking now governs both For You and Following feeds

Since 2025 X has migrated its ranking pipeline to Grok (xAI's model), and by 2026 Grok reads every post, watches every video, and personalizes predictions in real time. Both the For You feed and the Following feed are now Grok-ranked (Following used to be reverse-chronological). Confirmed engagement weights: a reply the author replies to is worth ~150× a like, a retweet is worth ~20× a like, a bookmark is worth ~10× a like.

What this means for you

The big shift: bookmarks are now a strong ranking signal. Content people save (long threads, checklists, reference material) gets boosted. For healthcare/practice-owner accounts: reference threads (e.g. 'what to know about IOP admissions') and framework posts save-out at high rates and gain compounding reach. Straight promotional posts and daily quips underperform relative to save-worthy content.

Read the source ↗
January 2026MEDIUM IMPACT

X publishes the full Grok-powered feed algorithm on GitHub

X published the complete Grok-powered feed algorithm source code publicly at github.com/xai-org/x-algorithm on January 20, 2026. First time the entire ranking system has been visible in this form. Confirmed that Grok reads and scores every post (text + images + video), personalizes in real time, and factors both engagement history and content-quality signals.

What this means for you

Practical takeaway from the source code release: originality signals matter (Grok can detect near-duplicate content), video content is weighted heavily (Grok watches it, not just captions), and account-quality signals (post frequency, follower ratio, complete profile) still gate reach. Nothing dramatic to change day-to-day, but confirms that quality-focused strategies outperform volume plays.

Read the source ↗

Frequently asked questions about X (Twitter) algorithm updates

How does Grok rank posts on X in 2026?

Grok reads every post's text, watches every attached video, and personalizes predictions in real time based on user history and engagement. Both the For You and Following feeds are Grok-ranked (the Following feed used to be reverse-chronological). The system heavily weights engagement quality: a reply that gets a reply from the author is worth roughly 150× a like, a retweet is worth ~20× a like, a bookmark is worth ~10× a like.

What engagement is worth the most on X in 2026?

Ranked by weight: replies that spark further conversation (150× a like), retweets (20×), bookmarks (10×), likes (baseline). Bookmarks are notably strong in 2026, they signal that content is worth saving for later reference, which Grok interprets as high-value. For practice owners, reference threads and framework posts (things people bookmark for later) compound reach better than one-off promotional posts.

Why do bookmarks matter so much on X now?

Grok interprets bookmarks as a strong 'save-worthy' signal, content valuable enough that the viewer wants to return to it. For healthcare and behavioral health accounts, this shifts strategy: reference threads (e.g. 'what to know about IOP admissions'), symptom-to-decision flowcharts, and framework posts save-out at high rates and get pushed to broader audiences. Straight promotional posts and daily quips underperform relative to reference-quality content.

How does the 'Following' feed differ from 'For You' now?

Both are Grok-ranked. The Following feed is filtered to accounts you follow but ordered by predicted engagement, not chronology. The For You feed pulls from the entire graph. In July 2026 X added a mutual-follower boost that surfaces more posts from accounts you follow who also follow you back, especially in reply threads. For B2B practice owners: replying to peers who follow you back builds compounding visibility.

Should I care about the Grok algorithm source code publication?

For practitioners, mostly not directly, the code is dense and evolving. But the release confirms things we can act on: originality signals matter (Grok detects near-duplicates), video content is weighted heavily (Grok watches it, not just captions), and account-quality signals (post frequency, follower ratio, complete profile) still gate reach. The public release removes ambiguity from what's a ranking signal and what isn't.

Where does Allgood Marketing get X algorithm data?

We track blog.x.com, X Engineering, the public github.com/xai-org/x-algorithm repository, Nikita Bier's product updates, and industry-side coverage from Social Media Today and Sprout Social. New confirmed X changes are appended to this timeline every 3 days.

Authoritative sources we track

Certified & Trusted

Google Partner Meta Business Partner Microsoft Advertising Partner Semrush Certified Salesforce Partner NVIDIA Partner

HIPAA & 42 CFR Part 2 Compliance

Allgood Marketing handles all client and patient data in alignment with HIPAA and 42 CFR Part 2. We sign Business Associate Agreements (BAAs) with covered-entity clients, avoid non-compliant pixels and forms on regulated pages, and configure ad accounts, analytics, and CRMs to current HHS OCR and FTC guidance. Read our full HIPAA & 42 CFR Part 2 policy or our Privacy Policy.

Made with ❤️ by ALLGOOD MARKETING

© 2026 Allgood Marketing™. All rights reserved.
ALLGOOD™ is a trademark of Allgood Marketing, LLC.

Book a Strategy Call