YouTube Data API adds videos.batchGetStats, charged to its own 10,000-unit quota bucket

Announced3 Jun 2026

Effective3 Jun 2026

StatusIn effect

What changed

YouTube's June 3, 2026 Data API revision adds videos.batchGetStats, which returns statistics for a list of videos. A call costs 1 quota unit in its own granular quota bucket, with a default of 10,000 units a day. The method returns viewCount, likeCount and commentCount, the publish time and the duration, plus a summary of how many videos succeeded or failed, because the call is intentionally not atomic. Authorization is not required for public videos. The quota cost table and the getting-started quota note do not list the new bucket yet.

What it means

Tools that refresh view, like and comment counts can read them with videos.batchGetStats, which draws on its own daily bucket instead of the shared 10,000-unit pool used by videos.list and most other methods.

Who is affected

Scheduling and analytics tools that poll YouTube video statistics for many videos, and the teams whose dashboards show those numbers.

What to do by 3 Jun 2026

Optional. Move bulk statistics reads from videos.list to videos.batchGetStats (up to 50 IDs per call), and handle partial results using the response summary.

What this means for you

Apps that show how many views and likes your YouTube videos have must ask YouTube for those numbers, and YouTube gives each app a daily allowance of questions. There is now a separate way to ask only for the numbers, with its own allowance. So your tool can check your video stats more often without running out of room for other jobs, like reading comments.

The details

Method: GET https://www.googleapis.com/youtube/v3/videos:batchGetStats, added June 3, 2026.
Quota: "A call to this method has a quota cost of 1 unit in its own granular quota bucket. The method's default quota is 10,000 units per day." This is separate from the shared bucket that videos.list (also 1 unit) draws on. See our quota buckets entry for the videos.insert and search.list buckets.
Parameters: id (required, comma-separated, at most 50 IDs per the discovery document), part (required: contentDetails, id, snippet, statistics) and onBehalfOfContentOwner (content partners only).
Response: items[] of videoStat resources with statistics.viewCount, likeCount, commentCount; snippet.publishTime; contentDetails.duration and durationMillis. A summary object reports requestedVideoCount, succeededVideoCount and failedVideoCount.
Not atomic: Google says the call "is intentionally not atomic", so some IDs can fail while others succeed.
Auth: not required for public videos. For non-public videos use one of youtube.readonly, youtube, youtube.force-ssl or youtubepartner.
Not stated: the quota cost table and the getting-started page still describe only the search.list and videos.insert buckets; neither lists batchGetStats yet.

What to do next

Business owners: Nothing.
Marketers and agencies: If your reporting tool hits YouTube quota limits, ask whether it uses the new stats method.
Developers: Read counts with videos.batchGetStats in batches of 50, check summary.failedVideoCount and retry the failed IDs, and keep videos.list for fields this method does not return.

Follow new entries with the RSS feed. Each platform also has its own feed, for example X or Meta.