Einstein Discovery Capacities and Requirements

Einstein Discovery has limits for CRM Analytics datasets, models, and predictions.

Note Einstein Discovery stories are now models. We wish we could snap our fingers to update the name everywhere, but you can expect to see the previous name in a few places until we replace it.

CRM Analytics Dataset Requirements

CRM Analytics datasets must meet these requirements.

Requirement	Minimum	Maximum
Number of rows	400	20,000,000
Number of columns	3 (1 outcome variable plus 2 other columns)	50

Monitor Usage

From Setup, in the Quick Find box, enter Usage, and then select Usage.

Model Limits

Einstein Discovery models have these limits.

Item	Limit
Models created or trained per org per day	100
Models created or trained per org per month. You can purchase more models in blocks of 1,000 with Einstein Analytics Plus - Additional Stories (1,000).	500
Concurrent model creation and training processes per org	2
Models that can be queued to run at one time	5

Prediction Definition Limits

Item	Limit
Active prediction definitions	CRM Analytics Plus license: Unlimited Einstein Prediction license: 10
Unique prediction definitions on a single entity	3
Active models deployed to a single prediction definition	10

Prediction Limits

Automated Prediction Writeback limits apply to scoring initiated through the Model Manager and scoring that is applied when records are created or updated.

Item	Limit
Predictions requested per org in a rolling 24-hour period using automated prediction fields.	5,000,000 Without high-volume prediction writeback, this limit is 500,000. High-volume prediction writeback is automatically enabled in orgs created after the Winter ’23 release. For orgs created before the Winter ‘23 release, it can be enabled. For instructions, see Save Higher Volumes of Predictions More Quickly (Generally Available).
Records that can be scored per org in a rolling 24-hour period using Score Data on the Scoring tab	1,000,000

Item

Limit

Predictions requested per org in a rolling 24-hour period using automated prediction fields.

5,000,000

Without high-volume prediction writeback, this limit is 500,000. High-volume prediction writeback is automatically enabled in orgs created after the Winter ’23 release. For orgs created before the Winter ‘23 release, it can be enabled. For instructions, see Save Higher Volumes of Predictions More Quickly (Generally Available).

Records that can be scored per org in a rolling 24-hour period using Score Data on the Scoring tab

1,000,000

On-demand limits apply to predictions initiated through Einstein Prediction Service, Lightning Components, Flow Builder, the PREDICT() function, Apex, Slack, or Einstein Discovery in Tableau.

Item	Limit
On-demand prediction requests per org in a rolling 24-hour period	50,000
On-demand prediction requests per user per hour	2,000
Concurrent on-demand prediction requests within an org	Depends on how the model associated with the request is built. Einstein Discovery built models: 25 Externally built models uploaded to Salesforce: 1

Item

Limit

On-demand prediction requests per org in a rolling 24-hour period

50,000

On-demand prediction requests per user per hour

2,000

Concurrent on-demand prediction requests within an org

Depends on how the model associated with the request is built.

Einstein Discovery built models: 25
Externally built models uploaded to Salesforce: 1

Prediction usage in recipes and dataflows isn’t limited, but Data Prep is subject to CRM Analytics Limits.

Other Limitations

Multivalue fields aren’t supported in Einstein Discovery. To learn more, see Limitations of Multivalue Fields in CRM Analytics.