Open data
7,669 accounts measured · CC BY 4.0We measured likes and comments per post against followers on 7,669 Instagram accounts. The median account under 10,000 followers engages 19.3% of them on a post; the median account above 10 million, 1.0%. Every band with its 25th, 50th, 75th and 90th percentile. Free data, CC BY 4.0.
Free forever plan · No credit card required · Cancel anytime
Likes plus comments per post, divided by followers, as percentiles across every measured account in the band. The bar is the median.
| Followers | Accounts | p25 | Median | p75 | p90 | Posts per account |
|---|---|---|---|---|---|---|
| Under 10K | 224 | 6.5% | 19.3% | 74.3% | 301.8% | 11.8 |
| 10K-100K | 1,214 | 1.6% | 6.4% | 17.9% | 48.5% | 11.7 |
| 100K-1M | 3,260 | 1.2% | 3.5% | 9.1% | 20.6% | 11.9 |
| 1M-10M | 2,423 | 0.6% | 2.0% | 5.2% | 11.2% | 12.0 |
| 10M+ | 548 | 0.3% | 1.0% | 2.5% | 5.5% | 11.9 |
Every band clears the API’s reliability threshold of 100 accounts. Upper percentiles can pass 100%: an account whose posts travel far beyond its followers collects more likes than it has followers, and that is a real measurement, not an error.
The rate divides likes and comments by followers. An account with 2,000 followers whose reel is pushed to 50,000 people by Instagram’s recommendation feeds collects more likes than it has followers, and its rate passes 100%. In the under-10K band the 90th percentile is 301.8%. That is what reach outrunning a follower count looks like in this metric, which is why we lead with the median and print the quartiles next to it.
The catalogue is built for discovering and monitoring accounts worth watching, not by sampling Instagram at random. Of the 7,684 accounts with a follower count, only 225 have fewer than 10,000, and most of those arrived through the related-profile edges of large accounts or because a post travelled. A typical small business page with 40 likes on 3,000 followers is under-represented here, so 19.3% describes small accounts that were already worth cataloguing, not small accounts in general.
This rate counts likes and comments only, over the posts the crawl holds (exactly 12 for 97.6% of accounts), and divides by followers rather than by reach. Pages that divide by reach, include saves and shares, or average a whole year of brand posts land somewhere else. The Limits section sets our figures next to theirs.
The median rate is 19.3% under 10,000 followers, 6.4% at 10K-100K, 3.5% at 100K-1M, 2.0% at 1M-10M and 1.0% above 10 million. The first step cuts it by two thirds; every later step roughly halves it. From the smallest band to the largest the median falls nineteenfold, and the decline is monotonic across all five bands, which is the shape the field expects.
At 100K-1M followers the bottom quarter of accounts sits below 1.2% and the top quarter above 9.1%: a sevenfold range inside one band, more than the gap between neighbouring medians. The same 4% is a shade above the median for an account with 500,000 followers, inside the top half but not the top quarter at 1M-10M, and inside the top quarter above 10 million. Read a rate against its band, never against a platform average.
Across the 21,664 catalogue accounts with both dates, 11.9% had not posted for 30 or more days on the day the crawl read them, 4.9% for 90 or more and 1.2% for more than a year. The median account had posted 2.3 days before its crawl. Smaller accounts go quiet more often: 17.4% of accounts under 10,000 followers were 30 days silent against 10.1% above 10 million.
Across the 1,464 sampled accounts with a measurable rhythm, the median account posts 0.7 times a week; 57.6% post less than weekly and 10.2% daily or more. The rhythm barely moves with size until the very top: accounts under 10K, at 10K-100K, 100K-1M and 1M-10M all sit between 0.56 and 0.66 posts a week, while accounts above 10 million post 1.27 a week and only 42.9% of them post less than weekly.
Reels make up 44.4% of the 20,454 posts in the sample and 86.6% of accounts post them; carousels are 21.5% of posts and appear on 46.0% of accounts. On the 1,028 accounts that post two or more formats, a reel earns 14.4% fewer likes and comments than the account’s own average at the median, a carousel 10.5% more and a single image about the same. Head to head, carousels beat reels on 60.6% of the 467 accounts posting both, and images beat reels on 55.6% of 730. The gap widens with size: reels out-earn carousels on 44.6% of the 83 accounts under 10K that post both, and on 24.0% of the 75 above 10 million. Views are not part of this rate, so the finding is that reels get watched and carousels get liked, not that reels fail.
Share of accounts that had not posted for 30 or 90 days on the day the crawl read them, for every catalogue row with both dates.
| Followers | In catalogue | Quiet 30+ days | Quiet 90+ days |
|---|---|---|---|
| Under 10K | 225 | 17.4% | 8.0% |
| 10K-100K | 1,222 | 17.2% | 7.9% |
| 100K-1M | 3,261 | 12.5% | 4.4% |
| 1M-10M | 2,428 | 9.6% | 2.7% |
| 10M+ | 548 | 10.1% | 2.8% |
| No follower count | 16,417 | 11.7% | 5.1% |
| All accounts | 24,101 | 11.9% | 4.9% |
Sample size: 21,664
Posts per week over the most recent posts the crawl holds, per follower band. Daily means seven or more a week; under weekly means fewer than one.
| Followers | Median posts / week | Daily or more | Under weekly |
|---|---|---|---|
| Under 10K | 0.66 | 9.4% | 57.4% |
| 10K-100K | 0.57 | 9.9% | 63.4% |
| 100K-1M | 0.56 | 9.7% | 62.5% |
| 1M-10M | 0.59 | 8.4% | 63.6% |
| 10M+ | 1.27 | 12.6% | 42.9% |
| No follower count | 0.76 | 11.4% | 55.7% |
| All accounts | 0.70 | 10.2% | 57.6% |
Sample size: 1,464
Formats of the posts in the sample, and how each format performs against the account’s own average interactions. A format is compared only on accounts that post at least two formats, so a reels-only account does not score its reels against themselves.
| Format | Share of posts | Accounts using it | Median vs account average | Beats the account average | Accounts compared |
|---|---|---|---|---|---|
| Reels | 44.4% | 86.6% | -14.4% | 38.2% | 950 |
| Carousels | 21.5% | 46.0% | +10.5% | 58.5% | 545 |
| Single images | 34.2% | 78.5% | +0.7% | 50.7% | 808 |
Best format on accounts with two or more rated formats (1,028): Reels 34.0% · Carousels 28.7% · Single images 37.3%
Head to head: on accounts posting both, carousels out-earn reels on 60.6% of 467 accounts, and out-earn single images on 63.7% of 325 accounts.
Sample size: 1,477 accounts, 20,454 posts
The catalogue does not represent the whole of Instagram. It is a set of accounts collected for discovery and monitoring, and it leans toward creators and public figures: of the 3,751 profile-complete accounts that carry an Instagram category label, 58.6% are Artist, Digital creator, Actor, Reel creator, Athlete or Public figure. Brand and company accounts are a minority here.
That population is the main reason our medians sit above the figures on most benchmark pages. Rival IQ’s 2026 report puts the all-industry median at 0.30% and Socialinsider’s at 0.48%, both measured on brand pages over a full year; Sprout Social quotes 2.19% for creators under 10,000 followers; Later gives 4-8% for accounts under 10K and 0.5-1.5% above a million, dividing by reach and counting saves and shares. Our rate counts likes and comments only, divides by followers, uses the 12 most recent posts rather than a year, and reports the median of a creator-heavy population. Different question, different number; neither is wrong.
The follower counts behind the ladder were read between 29 August and 19 September 2026, so the bands are a snapshot of that window rather than a long-run average, and the rate of an account whose recent posts happened to travel is a snapshot too.
Two thirds of the catalogue (16,417 of 24,101 accounts) has no follower count and is outside the ladder. Those accounts are inside the silence, cadence and format figures, which need timestamps and post types rather than followers.
The under-10K band holds 224 accounts. That clears the API’s reliability threshold of 100 but it is the thinnest band, and its upper percentiles are dominated by accounts whose posts reached far beyond their followers. Lead with the median.
Posting rhythm is computed over the most recent posts the crawl holds, 12 for almost every account, so posts per week describes a recent stretch and not a year. Format comparisons are within-account and count likes and comments; views are not part of the rate.
The API’s facets endpoint estimates language, category and follower distributions from a block sample of about 220 rows. We ship those in the JSON labelled as the sample they are and do not chart them; the band composition on this page comes from the full listing walk instead.
The data is CC BY 4.0. Use it, quote it, republish it. The only ask is attribution.
<a href="https://pinlyx.com/research/instagram-engagement-benchmarks-2026">Instagram engagement rate benchmarks 2026</a>, Pinlyx, CC BY 4.0The catalogue behind these figures, queryable account by account: measured engagement, cohort percentiles, cadence and content mix.
Turn the DMs and comments those accounts generate into tracked contacts in one inbox.
Our other social dataset: how many subscribers actually see a Telegram post, across 3.2 million channels.
Ready to ship
The Pinlyx Data API returns the same measured rate and cohort percentiles for any catalogued Instagram account, and the CRM turns the accounts you pick into contacts.
We value your privacy
We use cookies to improve our site, analyze traffic, and personalize ads. You can accept all, reject non-essential, or customize your choices. Read our Cookie Policy.