Estimated monthly

Only pay for what you use

You’ll only pay for data that you upload and access in each product (referred to as Labelbox Units or LBUs), and your unit cost decreases as your volume increases. Move the slider below to choose how much data you plan on uploading and using in each product to estimate your monthly LBUs. Using higher volumes? Contact sales for committed-use discounts. 

How much data do you have?

5k Rows

% to be labeled


% to be used in model


Need help with labeling?

On-demand premium labeling services starting at $8/hour, only pay for the time when labeling occursMeasured by a labeler's screen activity time


Small teams and individuals looking to validate early-stage experiments.

Limit of 10K LBUs

  • Catalog

  • Annotate

  • Model

  • Customer support


Companies developing a few ML models. Discount available for annual commitment.

Everything in free plus:

  • Additional workspaces

  • Advanced analytics

  • Prioritized customer support


Companies managing multiple ML initiatives across various teams. Discount available for annual commitment.

Everything in Free and Standard plus:

  • Custom contract

  • Single sign-on

  • Custom editors


Data row type










1 LBU per data row

1 LBU per data row

1 LBU per data row


1 LBU per data row + 1 LBU per 10 frames

1 LBU per labeled frame (keyframe)

1 LBU per data row + 1 LBU per 10 frames

Geospatial & Medical Tiled Imagery


10 LBU per data row

10 LBU per labeled data row

1 LBU per labeled slice

10 LBU per data row

Compare features





Data volume

Determined by the number of LBUs (a normalized unit of data in the Labelbox Platform)

Limit of 10k



Unlimited querying and slices

To search & filter for data rows in Catalog

Precomputed and custom embeddings

Use embeddings to query and explore your unstructured data

Similarity search

Find similar data with embeddings

Custom metadata

Add metadata to tag a set of data rows

Annotation & metadata metrics

Distributions to better understand your data


Data Types





Label native PDF documents with NER and OCR techniques

Geospatial maps

Label simple & ESPG coordinate system

DeepZoom maps

Label pathology and large format TIFF files


Conversational Text

Label conversation or thread-based text data


Label HTML files intended to be viewed in a specific manner

Medical data

Label multi-frame DICOM files


Develop custom labeling interfaces with the Python SDK

Quality Control & Assurance

Custom labeling & review workflows

Create a highly-customizable review pipeline with up to 10 custom tasks

Issues & Comments

Create issues on an asset and facilitate collaboration with comments

Role-based access controls

Set permission levels depending on a user's role


Model-assisted labeling

Import pre-labels from your model or a model of your choice on an asset

Auto-labeling tools

Generate mask predictions with our image auto-segment tool

Team & Project Management

Project Performance Dashboard

Project-level and individual labeler analytics on throughput, efficiency, and quality

Advanced Analytics

A global view of all of your labeling operations and activity across all of your Labelbox projects


10 seats included

Multiple workspaces

Create more than one workspace to manage multiple business units or teams




Custom embeddings

Custom embeddings to query and explore your unstructured data

Model metrics (F1, IoU, precision, recall, and more)

Use our built-in metrics or upload your own custom metrics to measure the quality of your model

Unlimited querying

To search & filter for data rows in Model

Model training service integration

Train a model with your labeled data in one click

Data & model parameters versioning

Create, version, and track your hyperparameters and any training-related configurations for a model run

Render annotation and predictions

Visually inspect annotations and predictions

Error analysis workflows

Analyze where model predictions and ground truth labels disagree

Active learning workflows

Surface and prioritize high-value data to label to improve model performance

Security & Support

Single sign-on

We support most SSO identity providers including SAML, Okta, OneLogin, Google Apps, ADFS, and Azure


For customers with healthcare-related data, a HIPAA compliance program is critical.

Support available

Work with the Support Team to troubleshoot issues and for ongoing support on Labelbox best practices

Customer support

Prioritized customer support

Prioritized customer support

Free for research & education

Labelbox is free to use for those who attend and teach at qualified educational institutions to use for non-commercial research.

Learn more

Frequently asked questions

  • What is an LBU?

    A Labelbox Unit (LBU) is a normalized unit of data in the Labelbox Platform. An LBU is equivalent to one data row in each Labelbox product: Catalog, Annotate, and Model. 

    1. One image data row in Catalog is equal to one LBU 

    2. One image data row labeled in Annotate is equal to one LBU – no matter how many labels or annotations are created, deleted, or re-labeled on that one data row 

    3. One image data row in Model is equal to one LBU – no matter how many model runs that data row participates in

    More complex data rows can vary in LBU consumption, please refer to the table below to understand how LBU consumption varies based on data row type.

  • How does Labelbox calculate LBUs?

    You are charged based on the number of LBUs your team accrues monthly. As your LBU count increases, the monthly cost per LBU decreases. The more you use Labelbox and derive value across Catalog, Annotate, and Model, the less you pay per additional LBU. Please refer to our pricing calculator above to estimate your monthly cost.

  • What happens if I delete data? How am I charged?

    If you delete data rows from Catalog, those data rows will be removed from your total LBU count and you won’t be charged for those data rows in next month’s billing cycle. 

    Note: Deleting data rows in Catalog will prevent usage of those data rows in Annotate and Model. 

  • If I use the same data row (with the same data row ID) in two different model runs, but within the same model, am I charged 1 LBU or 2 LBUs?

    You will be charged 1 LBU for that data row in Model, regardless of the number of model runs that data row is used in. You can even use that data row across different models and you will only be charged for 1 LBU in our Model product. 

    Learn more about LBU accrual in specific scenarios across Catalog, Annotate, and Model in our documentation.

  • What happens if I hit my Free tier limit?

    The current Free tier limit is 10k LBUs. If you hit your Free Tier limit, but want to continue using Labelbox for your current projects, you can upgrade to a Standard Labelbox plan. Learn more about upgrading your plan in our documentation.

  • How do I know if Standard or Enterprise is right for me?

    A unique aspect of our pricing is that you only pay for what you need. We offer unlimited querying, searching, and annotations for both tiers.

    Pricing is determined by how much data you anticipate you’ll want to store or annotate, and ultimately how fast you want to develop your model. We recommend our Enterprise tier for teams who want to scale faster with greater security & compliance measures, more collaborative features, and best-in-class support. Contact sales to learn more. 

  • Does Labelbox provide outsourced labeling?

    Yes – with Labelbox Boost, we offer on-demand premium labeling services designed for your use case and offer expert guidance every step of the way. 

    Our suite of tools are designed to reduce the time it takes to label each asset, so our labelers are focused on completing tasks faster and on increasing the value of each label. Rather than paying for labeling work on a fixed label-cost model, you only pay for the time when labeling occurs. 

  • How does Labelbox help me spend less time and cost labeling?

    We believe that the best way to spend less on labeling is to automate your labeling process to accelerate efforts without diminishing the quality of annotations.

    Labelbox provides several automation techniques, most notably pre-labeling or model-assisted labeling, to decrease labeling costs and help your model get smarter with every iteration. Teams who use model-assisted labeling at Labelbox have seen a reduction in labeling time and cost by 50-70%.

  • How secure is my data with Labelbox?

    We believe that your data should stay your data. We don’t use customer data in any tier and go even further by allowing everyone to integrate their cloud buckets via IAM Delegated Access to keep your raw data in your own cloud storage service.

    If you’re interested in additional security measures, we offer Enterprise-grade security for our Enterprise tier, including SSO, HIPAA, Learn more about Security at Labelbox