Data study · 10 min read
Best time to post on X (Twitter) in 2026
The best time to post on X (Twitter) is 6am to 9am Eastern Time, and the best day is Sunday. In 289,640 original English posts from 4,399 creators (Jan–Oct 2026), Sunday posts went viral (5x the creator's median views) 13% more often than those creators' usual rate, and posts sent from 6am to 9am ET about 6% more often.
A link (−24%) or video (+19%) moved the odds more than any single hour (−7% to +8%).
Joschua Sünneke·$1 for 7 days · Cancel anytime
- Original X posts
- 289,640
- Creators
- 4,399
- Went viral
- 30,127
- 5x the creator's median views
- Posted
- Jan–Oct 2026
Source: Postel's viral post library, original English X posts, aggregated.Download the data (CSV)
Key findings
- Sunday is the best day to post on X, Wednesday the worst. Sunday +13% (95% range +10% to +16%), Saturday +6%, Wednesday −6% (the other ranges are in the chart below). These are relative changes, not percentage points: at the sample's 10.4% viral rate, +13% would mean about 11.7% of posts going viral instead of 10.4%.
- Sunday morning is the strongest stretch of the week. Posts sent on Sunday between 6am and noon ET went viral about 21% more often than usual. That figure comes from two 3-hour blocks without a range of their own, so the exact size is uncertain.
- The best hours are early: 6am to 9am Eastern. Across the week, posts sent then went viral about 6% more often than usual; on weekdays alone, about 4% more. 8am ET was the best hour (+8%, 95% range +2% to +13%) and 2pm ET the weakest (−7%, 95% range −10% to −3%).
- The busiest hours aren't the best ones. 27% of posts go out between 9am and 1pm ET, the four busiest hours, but together those hours come out close to usual (about +2%). Of those four, only 9am is clearly above usual (+5%).
- Longer posts went viral more often. Posts of 501 to 1,000 characters: +20% (95% range +16% to +23%). Posts of 50 to 100 characters: −9% (95% range −11% to −7%).
- Video and multi-image posts beat a single image. Video +19%, 2 to 4 images +16%, one image −13%, a GIF −34%.
- Posts with a link went viral 24% less often (95% range −28% to −19%). 8.9% of posts carried one.
- Thread openers did best of any group, with a wide range. +61% (95% range +39% to +85%), from only 1,626 posts; the rule that finds them also catches some numbered lists.
The best time to post on X, by hour
Share of posts sent
Viral posts vs the creators' usual
Times are US Eastern (EDT for 91% of the sample). The creators post from many time zones, so the hours mark when posts went out; they say nothing about anyone's local time.
Show the numbers
| Hour (ET) | Share of posts | Viral posts vs usual | 95% range | Posts |
|---|---|---|---|---|
| 12am | 2.1% | +2.9%−3.6% to +9.7% | −3.6% to +9.7% | 6,000 |
| 1am | 2.3% | −5.6%−14.2% to +3.0% | −14.2% to +3.0% | 6,563 |
| 2am | 2.5% | −2.2%−7.8% to +4.3% | −7.8% to +4.3% | 7,305 |
| 3am | 2.7% | −3.1%−9.3% to +2.6% | −9.3% to +2.6% | 7,867 |
| 4am | 2.9% | +1.7%−4.3% to +7.5% | −4.3% to +7.5% | 8,269 |
| 5am | 3.1% | −1.0%−6.9% to +5.0% | −6.9% to +5.0% | 9,012 |
| 6am | 3.4% | +7.3%+1.2% to +13.4% | +1.2% to +13.4% | 9,775 |
| 7am | 4.2% | +4.5%−2.3% to +11.3% | −2.3% to +11.3% | 12,109 |
| 8am | 5.5% | +7.5%+2.1% to +13.3% | +2.1% to +13.3% | 16,001 |
| 9am | 6.5% | +4.9%+0.6% to +9.5% | +0.6% to +9.5% | 18,833 |
| 10am | 7.0% | +2.3%−2.0% to +6.7% | −2.0% to +6.7% | 20,253 |
| 11am | 7.1% | +0.6%−3.8% to +4.4% | −3.8% to +4.4% | 20,543 |
| 12pm | 6.8% | −0.1%−3.1% to +4.0% | −3.1% to +4.0% | 19,799 |
| 1pm | 6.4% | −3.1%−7.0% to +1.6% | −7.0% to +1.6% | 18,598 |
| 2pm | 6.0% | −6.6%−10.4% to −2.8% | −10.4% to −2.8% | 17,329 |
| 3pm | 5.4% | −1.2%−5.6% to +3.5% | −5.6% to +3.5% | 15,557 |
| 4pm | 4.8% | +2.7%−2.6% to +7.1% | −2.6% to +7.1% | 13,848 |
| 5pm | 4.3% | −2.6%−7.6% to +2.1% | −7.6% to +2.1% | 12,326 |
| 6pm | 3.7% | −2.1%−7.2% to +3.3% | −7.2% to +3.3% | 10,719 |
| 7pm | 3.3% | −2.8%−8.5% to +1.8% | −8.5% to +1.8% | 9,415 |
| 8pm | 2.9% | −3.4%−9.2% to +2.8% | −9.2% to +2.8% | 8,419 |
| 9pm | 2.7% | −4.9%−11.0% to +1.4% | −11.0% to +1.4% | 7,760 |
| 10pm | 2.4% | −3.7%−10.7% to +3.2% | −10.7% to +3.2% | 6,935 |
| 11pm | 2.2% | +0.8%−6.6% to +7.9% | −6.6% to +7.9% | 6,405 |
A raw count of viral posts per hour mostly shows when people post: the four busiest hours, 9am to 1pm ET, would come out on top because they hold the most posts. Compared with each creator's own usual, the early hours come out ahead instead. Hollow bars in the chart are hours whose range crosses zero: they aren't clearly different from usual.
Try your first post of the day between 6am and 9am ET. That is early morning on the East Coast and lunchtime in Europe, but 3am to 6am on the West Coast. If West Coast readers matter, 9am ET (6am Pacific) is the latest hour still clearly above usual: +5% (95% range +1% to +10%).
The best day to post on X
Show the numbers
| Day (ET) | Viral posts vs usual | 95% range | Posts |
|---|---|---|---|
| Monday | +2.2%−0.4% to +5.2% | −0.4% to +5.2% | 41,041 |
| Tuesday | −2.9%−5.2% to −0.1% | −5.2% to −0.1% | 43,760 |
| Wednesday | −6.4%−8.9% to −4.1% | −8.9% to −4.1% | 45,094 |
| Thursday | −4.3%−6.6% to −2.2% | −6.6% to −2.2% | 46,478 |
| Friday | −2.9%−5.4% to −0.2% | −5.4% to −0.2% | 45,291 |
| Saturday | +6.4%+3.5% to +9.6% | +3.5% to +9.6% | 35,463 |
| Sunday | +12.8%+9.6% to +15.6% | +9.6% to +15.6% | 32,513 |
Fewer posts go out at the weekend: 23% of the sample was posted on Saturday or Sunday, against two sevenths (29%) if posting were even. Fewer posts competing for the same readers may be one reason weekend posts travel further. Monday was about usual and Thursday close to Wednesday. If you only post on weekdays, adding one weekend post is the cheapest test in this study.
Best day and time together
Each cell holds 2,315 to 9,826 posts and has no 95% range of its own, so neighboring cells can differ by several percent by chance.
Show the numbers
| Day and time (ET) | Viral posts vs usual | Posts |
|---|---|---|
| MonMonday 12–3am | +8.7% | 2,796 |
| MonMonday 3–6am | +9.0% | 3,574 |
| MonMonday 6–9am | +9.1% | 5,375 |
| MonMonday 9am–12pm | +4.2% | 8,399 |
| MonMonday 12–3pm | −3.1% | 7,908 |
| MonMonday 3–6pm | −0.7% | 5,877 |
| MonMonday 6–9pm | −4.4% | 4,111 |
| MonMonday 9pm–12am | +1.2% | 3,001 |
| TueTuesday 12–3am | −7.5% | 2,827 |
| TueTuesday 3–6am | −2.1% | 3,607 |
| TueTuesday 6–9am | +5.0% | 5,610 |
| TueTuesday 9am–12pm | +1.2% | 8,964 |
| TueTuesday 12–3pm | −4.5% | 8,624 |
| TueTuesday 3–6pm | −2.4% | 6,496 |
| TueTuesday 6–9pm | −10.1% | 4,399 |
| TueTuesday 9pm–12am | −9.6% | 3,233 |
| WedWednesday 12–3am | −9.8% | 3,046 |
| WedWednesday 3–6am | +0.2% | 3,756 |
| WedWednesday 6–9am | −0.9% | 5,868 |
| WedWednesday 9am–12pm | −2.8% | 9,372 |
| WedWednesday 12–3pm | −11.0% | 8,619 |
| WedWednesday 3–6pm | −8.5% | 6,719 |
| WedWednesday 6–9pm | −8.0% | 4,485 |
| WedWednesday 9pm–12am | −11.4% | 3,229 |
| ThuThursday 12–3am | −8.0% | 3,050 |
| ThuThursday 3–6am | −8.2% | 4,067 |
| ThuThursday 6–9am | +0.6% | 5,951 |
| ThuThursday 9am–12pm | −1.4% | 9,826 |
| ThuThursday 12–3pm | −8.7% | 9,001 |
| ThuThursday 3–6pm | −5.6% | 6,654 |
| ThuThursday 6–9pm | −2.1% | 4,618 |
| ThuThursday 9pm–12am | −2.4% | 3,311 |
| FriFriday 12–3am | −0.2% | 3,108 |
| FriFriday 3–6am | −14.2% | 4,022 |
| FriFriday 6–9am | +5.6% | 6,193 |
| FriFriday 9am–12pm | −3.7% | 9,625 |
| FriFriday 12–3pm | −6.4% | 8,690 |
| FriFriday 3–6pm | 0% | 6,504 |
| FriFriday 6–9pm | +2.6% | 4,160 |
| FriFriday 9pm–12am | −9.9% | 2,989 |
| SatSaturday 12–3am | −5.4% | 2,726 |
| SatSaturday 3–6am | −0.6% | 3,365 |
| SatSaturday 6–9am | +10.4% | 4,813 |
| SatSaturday 9am–12pm | +6.6% | 7,097 |
| SatSaturday 12–3pm | +15.2% | 6,647 |
| SatSaturday 3–6pm | +4.8% | 4,909 |
| SatSaturday 6–9pm | +0.7% | 3,325 |
| SatSaturday 9pm–12am | +8.3% | 2,581 |
| SunSunday 12–3am | +13.3% | 2,315 |
| SunSunday 3–6am | +15.9% | 2,757 |
| SunSunday 6–9am | +21.0% | 4,075 |
| SunSunday 9am–12pm | +20.3% | 6,346 |
| SunSunday 12–3pm | +2.8% | 6,237 |
| SunSunday 3–6pm | +16.3% | 4,572 |
| SunSunday 6–9pm | +4.9% | 3,455 |
| SunSunday 9pm–12am | +7.3% | 2,756 |
The four strongest blocks are all on Sunday (Saturday 12–3pm is next, at +15%). Pooled across Monday to Friday, 6am to 9am is the only block above usual (about +4%); single weekday cells, like Monday before 9am, can run higher. Weekday afternoons and evenings mostly sit below usual, Wednesday most of all. The blocks have no ranges of their own, and they are noisy: Sunday noon to 3pm was only +3% while the blocks either side of it were higher. Read the grid as Sunday, mostly the morning, not as a ranking of single blocks.
What matters more than the time you post
By hour, results ranged from −7% to +8%. What's in the post ranged far wider, from a GIF (−34%) to a thread opener (+61%). For the rest of what drives reach, see how to get more impressions on X.
Length
Posts under 50 characters are not in the dataset, so the shortest bucket starts at 50.
Show the numbers
| Length | Viral posts vs usual | 95% range | Posts |
|---|---|---|---|
| 50–100 characters | −9.2%−10.7% to −7.3% | −10.7% to −7.3% | 80,186 |
| 101–140 characters | −6.9%−9.7% to −4.2% | −9.7% to −4.2% | 39,828 |
| 141–200 characters | −5.5%−8.4% to −2.4% | −8.4% to −2.4% | 39,191 |
| 201–280 characters | +0.2%−2.4% to +2.6% | −2.4% to +2.6% | 46,743 |
| 281–500 characters | +10.3%+7.1% to +13.1% | +7.1% to +13.1% | 41,169 |
| 501–1,000 characters | +19.7%+16.3% to +23.2% | +16.3% to +23.2% | 24,286 |
| 1,001–2,500 characters | +16.9%+12.1% to +22.6% | +12.1% to +22.6% | 15,014 |
| 2,500+ characters | +30.9%+19.2% to +42.7% | +19.2% to +42.7% | 3,223 |
Because the comparison is within each creator, this isn't just Premium accounts doing better: creators' own long posts beat their own short ones. Posts over 2,500 characters did best, but there were only 3,223 of them, so their range is wide (95% range +19% to +43%).
Media and links
Media
Links
Threads
A thread opener has a 🧵, starts with “1/”, “1.” or “1)” (so some numbered lists count), or says “thread” followed by a colon or a down arrow. Replies aren't in the dataset, so the rest of a thread isn't counted.
Show the numbers
| Post type | Viral posts vs usual | 95% range | Posts |
|---|---|---|---|
| Video | +19.0%+16.5% to +21.1% | +16.5% to +21.1% | 58,606 |
| 2–4 images | +15.7%+11.8% to +21.1% | +11.8% to +21.1% | 19,448 |
| Text only | −3.0%−4.5% to −1.4% | −4.5% to −1.4% | 113,754 |
| 1 image | −13.4%−15.2% to −11.4% | −15.2% to −11.4% | 96,036 |
| GIF | −34.0%−50.4% to −20.4% | −50.4% to −20.4% | 1,796 |
| No link | +1.9%+1.5% to +2.3% | +1.5% to +2.3% | 263,831 |
| Has a link | −23.5%−28.1% to −19.1% | −28.1% to −19.1% | 25,809 |
| Thread opener | +61.1%+39.1% to +85.2% | +39.1% to +85.2% | 1,626 |
| Single post | −0.3%−0.5% to −0.2% | −0.5% to −0.2% | 288,014 |
Text-only posts were slightly below usual (−3%, 95% range −5% to −1%). GIFs did worst, though only 1,796 posts used one. Posts with a link had lower median views as well as fewer viral posts. If a link matters, put it in a reply under the post: Postel can add it as a first comment when the post goes out. That reply still counts as an X post with a link, which most Postel plans cap each month.
Thread openers
Posts marked as a thread went viral 61% more often than usual (how a post was counted as a thread opener is under the chart above). This is the smallest group in the study (1,626 posts, 95% range +39% to +85%), so it's worth a test, but the size is uncertain. The tweet thread maker writes and schedules threads.
How to find your own best time to post on X
This study averages 4,399 creators. Your audience keeps its own hours, so test them:
- Start from the data. Put one weekday post between 6am and 9am ET and one on Saturday or Sunday morning. Shift by your audience's time zone if it isn't mostly in the US.
- Set posting slots. Add those times as weekly slots in Postel's X scheduler, and each post you queue lands in the next free one. Plan next week can fill those slots with drafts for you to approve; other tools that do slots are in the best X schedulers.
- Post at a mix of times for 3 to 4 weeks. Aim for at least two posts a week in each slot, so each time has enough posts to compare.
- Compare posts with your own usual, not raw views. A post with 5x your median views went viral by this study's measure, whatever its total. On X, Postel's Analytics scores each post against your usual and shows your best days and hours once 8 of your posts have stats.
- Keep what beats your usual and drop the rest. Recheck every few months; audiences move.
Turn these times into your schedule
Add your best times as weekly slots and Postel posts your queue into them on X, LinkedIn, Bluesky and Mastodon. For You writes posts in your voice from what's going viral on your topics, ready for those slots.
8,000+ people have signed up
$1 for 7 days · Cancel anytime
Who this data fits
Best for: English-language creators and founders with an established audience who publish original posts several times a week. The median creator in the sample had 51,565 followers.
Not a fit for: accounts whose readers are all in one time zone outside the US (shift the hours, or test your own), very small accounts (posts under 400 views aren't in the dataset), replies and reposts (not studied), and other networks (this is X data only).
Methodology
- Source. Postel's viral post library: original posts from the X creators Postel tracks for its users. Replies, reposts and quote posts are filtered out on import. A post is stored once it has 400 views and 50 characters. The library held 497,697 posts from the window; 289,640 passed every filter below.
- Filters. English posts, posted January 1 to October 26, 2026. Views had at least 24 hours to settle (the post was first stored a day or more after it went out). Each creator needed at least 20 posts in the sample, so each usual rate rests on enough posts. That left 4,399 creators, with a median of 8,703 views per post.
- Viral. A post went viral when it got at least five times its creator's median views. 30,127 posts (10.4%) did.
- Compared with each creator's usual. "Usual" on this page always means the creator's own viral rate across all their posts, not their views. For a group of posts (an hour, a day, a length), we compare the viral posts it got with what those creators' rates predict. 0% means as expected; +10% means 10% more viral posts than those creators usually get, a relative change: at the sample's 10.4% rate it would mean about 11.4% of posts going viral instead of 10.4%. A creator who posts a lot at 9am only lifts 9am if their 9am posts beat their own usual.
- Time zone. X timestamps are UTC. We converted them to US Eastern Time, accounting for daylight saving. 91% of the sample was posted under EDT, more than the calendar share, because the library holds more posts from spring onward. The CSV also has the UTC hours.
- Uncertainty. The 95% ranges come from resampling creators 300 times. Groups whose range crosses 0% aren't clearly different from usual. The day-and-time blocks have no ranges.
- Checks. The pattern held when we required 0 or 48 hours of settling instead of 24, and when "viral" meant 3x or 10x the creator's median instead of 5x. In all nine versions, Sunday and Saturday stayed on top, Wednesday at the bottom, and 6am to 9am ET, pooled across the week, above usual (lowest +1%, with 48 hours of settling and the 10x cut).
Limitations
- X only. No data from other networks.
- Not a random sample of X. These are creators tracked because their posts tend to do well, and posts under 400 views aren't stored. Small accounts may see a different pattern.
- Time zones are unknown. Creators post from everywhere, so Eastern Time here only marks when posts went out; it says nothing about where readers were.
- Views are a snapshot. We read views when the post was stored or last refreshed. Posts kept collecting views after that, mostly in their first days.
- The last weeks may be thin. Collection of new posts slowed from about October 12, 2026, so the final two weeks of the window may hold fewer posts than earlier weeks.
- Correlation, not cause. Creators may save their best ideas for Sunday. The within-creator comparison removes who is posting, not what they chose to post when.
Download the data
The aggregates behind every chart are free to use under CC BY 4.0, with a link to this page: best-time-to-post-on-x-2026.csv. Each row is one group (an hour in ET or UTC, a day, a day-and-time block, a length, a media type, links or threads) with its post count, creators, viral posts, the viral posts expected from those creators, the difference in percent with its 95% range, and median views against the creators' usual. Plain columns (hour, day, block start, character range) make it easy to sort and join. No post text or creator names are included.
To cite this study
Sünneke, J. (2026, October 29). Best time to post on X (Twitter) in 2026 [Data set and analysis]. Postel. 289,640 posts from 4,399 creators. https://www.postel.app/best-time-to-post-on-x
Frequently asked questions
Updated October 29, 2026, with posts through October 26, 2026.
What is the best time to post on X (Twitter)?
6am to 9am Eastern Time, and the best day is Sunday. In our study of 289,640 posts from 4,399 creators, Sunday posts went viral 13% more often than the same creators' usual rate (95% range +10% to +16%). Across the week, 6am to 9am ET did best, about 6% more often than usual (about 4% more on weekdays alone).
What is the best day to post on X?
Sunday, then Saturday. Sunday posts went viral 13% more often than usual and Saturday posts 6% more. Wednesday was the worst day: its posts went viral 6% less often.
What is the worst time to post on X?
Weekday afternoons and evenings, with 2pm ET the weakest hour (−7%) and Wednesday the weakest day (−6%). The gaps are small: a link cost more (−24%) than the worst hour did.
Does the time you post on X matter?
A little. By hour, how often posts went viral ran from −7% (2pm ET) to +8% (8am ET) against each creator's own usual. A link (−24%), video (+19%) or a post of 501 to 1,000 characters (+20%) moved the odds more.
How long should a post on X be?
With X Premium, aim for 501 to 1,000 characters: those posts went viral 20% more often than usual (95% range +16% to +23%). Without Premium, use most of the 280: posts of 201 to 280 characters came out at usual, while posts of 50 to 100 characters went viral 9% less often. The few posts over 2,500 characters did best (+31%), with a wide range.
Do links hurt reach on X?
In this sample, yes. Posts with a link went viral 24% less often than usual for the same creators. Putting the link in a reply keeps the post itself link-free.
Which time zone are these times in?
US Eastern Time (EDT for most of the sample). The raw timestamps are UTC, and the CSV has both. The creators post from many time zones, so if your readers are in one place, shift the hours or test your own.
Can Postel post at my best times?
Yes. Set weekly posting slots once and every post you add to the queue goes into the next free slot, on X, LinkedIn, Bluesky and Mastodon. On X, Analytics shows the days and hours your own posts do best once 8 of them have stats. It doesn't change your slots on its own.
Can I use this data?
Yes. Download the CSV and use it with a link to this page. It holds aggregates only, with no post text or creator names.