Updated October 10, 2026. Data: OpenTweet, pulled Oct 10 2026.
Does X penalize links in 2026? What 28,124 posts show
Not in OpenTweet's data. Across 60 X accounts that posted at least five posts with a link and five without in 2026, the typical account's link posts got 1.05 times the impressions of its no-link posts, and links did worse on 27 of the 60. X said on July 29, 2026 that it does not penalize links.
Does X penalize links? Most answers online quote a percentage with no sample behind it. Below: what X said, what 28,124 posts published through OpenTweet show, how we counted, what the data cannot tell you, and the separate API cost that still makes links more expensive.
What X said
On July 29, 2026 Paul Graham asked Elon Musk whether X no longer penalizes posts with links. Musk replied:
"We haven't for over a year."
A few hours earlier, on July 28, Nikita Bier had posted "you do not need to put the links in replies anymore". Both posts were reported by PPC Land on July 30, 2026.
The same report cites PPC Land's own analysis of X's ranking model from January 2026, which found link posts got about 94% fewer views than comparable posts without them. So the official line and the outside estimates disagree. Our data is a third view, based on real posts and real impressions.
What OpenTweet's data shows
In OpenTweet's data, posts with links did not get fewer impressions than the same account's posts without links. The within-account ratio is the fairer number, because it compares each account only with itself.
| Measure | Value |
|---|---|
| Period | January 1 to October 10, 2026 |
| Posts with a link | 13,815 |
| Posts without a link | 14,309 |
| Median impressions, all link posts | 57 |
| Median impressions, all no-link posts | 44 |
| Accounts in the within-account comparison | 60 (at least 5 link and 5 no-link posts each) |
| Median ratio, link to no-link impressions | 1.05 |
| Accounts where links did worse | 27 of 60 |
| Since July 29, 2026: accounts | 58 |
| Since July 29, 2026: median ratio | 1.07 |
| Since July 29, 2026: accounts where links did worse | 25 of 58 |
Data: OpenTweet, pulled Oct 10 2026.
Median impressions, all posts pooled
Mixes accounts of different sizes, so read it with care.
Link vs no-link, same account
Median of each account's link to no-link impressions ratio.
baseline
links lower on 27
links lower on 25
How we counted
- Only posts published to X through OpenTweet that have synced impressions.
- Text posts and posts with a single media item. Threads are left out.
- A post is a link post if its text contains http://, https:// or t.co/. Everything else is a no-link post.
- Period: January 1 to October 10, 2026, with a second cut from July 29, 2026, the day of Musk's reply.
- For the within-account comparison we kept only accounts with at least five link posts and five no-link posts, took each account's ratio of link to no-link impressions, and report the median across accounts.
What this data cannot tell you
Read these before quoting the numbers
- Mostly small accounts. The median post gets about 50 impressions. Large accounts may behave differently.
- Observational, not causal. People choose when to add a link, and that choice can change the result. This is not a controlled test.
- Impressions only. We did not measure clicks, likes, replies or follows.
What the data does show: for the accounts we can see, adding a link did not come with lower impressions. It does not prove that X boosts links, and it does not rule out a penalty for some kinds of accounts or links.
The other reason links cost more: the X API
Reach is one question, price is another. Since April 20, 2026, any post created through the X API that contains a URL bills at $0.20, against $0.015 for a plain post. Tools that post for you through the API pay that too, which is why some schedulers strip links or charge extra for them.
Moving the link into a reply does not avoid it. A reply created through the API is itself a post, so your own reply carrying the link bills at the link rate too. The details are in the X API link post fee, and what schedulers do about it is in why links are removed from scheduled X posts.
What to do
- Put the link in the post if the post is about the link. Our data gives no reason to hide it in a reply, and a reply with a link still bills at the link rate through the API.
- Check the preview first. A broken card makes a link post look worse than it is. Run the URL through the X card validator.
- Test on your own account. Post a few with and a few without, and compare impressions over a couple of weeks. Your audience matters more than any average.
- Write the post so it works without the click. Score a draft with the X algorithm score tool before you schedule it.
- Watch for changes. X's ranking changes often. We track them on the X algorithm changes page, and the background is in how the X algorithm works and the open source X algorithm on GitHub.
7-day free trial. Cancel anytime.
Frequently asked questions
Does X penalize links in 2026?
Not in OpenTweet's data. Across 60 X accounts that posted at least five posts with a link and five without in 2026, the typical account's link posts got 1.05 times the impressions of its no-link posts, and links did worse on 27 of the 60. X said on July 29, 2026 that it does not penalize links.
What did X say about the link penalty?
On July 29, 2026 Paul Graham asked Elon Musk whether X no longer penalizes posts with links, and Musk replied "We haven't for over a year." A few hours earlier, on July 28, Nikita Bier posted "you do not need to put the links in replies anymore".
Do tweets with links get less reach?
Not in OpenTweet's data. Comparing each account with itself, the median account's link posts got 1.05 times the impressions of its posts without links, from January 1 to October 10, 2026. Since July 29 the figure is 1.07 across 58 accounts. This is observational data from mostly small accounts, so it does not prove links help or hurt.
Should I put the link in the reply instead?
For reach, our data gives no reason to. For cost, it does not help either if you post through the X API: a reply created through the API is itself a post, so your own reply carrying the link bills at the link rate, $0.20, just like a post with a URL.
Why do other sites say links get far less reach?
Many of those figures do not publish a sample size or method, and some predate X's July 2026 statement. PPC Land reported about 94% fewer views for link posts based on its analysis of X's ranking model from January 2026. Our numbers come from real posts with synced impressions, compared within the same account.
Where does this data come from?
From posts published through OpenTweet between January 1 and October 10, 2026 that have synced X impressions. Text and single-media posts are included, threads are excluded, and a post counts as a link post if its text contains http://, https:// or t.co/. Data pulled October 10, 2026.
Keep exploring
Why are links removed from scheduled X posts?
Which schedulers strip links, which charge, and why.
The X API link post fee
What the $0.20 link price means for API posting.
How the X algorithm works in 2026
How posts are ranked in the For You feed.
X algorithm score
Score a draft post before you publish it.
X card validator
Check how your link preview will look on X.
X algorithm changes
Dated log of ranking changes on X.
Schedule link posts and see their impressions.
OpenTweet syncs impressions for the X posts it publishes, so you can compare your own link and no-link posts. 10 link posts a day on Advanced for $29, 20 on Agency for $49.
7-day free trial. Cancel anytime.