Looker vs Metabase

Quick verdict

Looker is Google Cloud's enterprise BI platform. Developers describe tables, joins, dimensions and measures in LookML files stored in Git; business users explore through the browser, and Looker writes SQL against the connected database for every query. It is sold as Looker (Google Cloud core), managed in the Google Cloud console, and as the older Looker (original), which can be Looker-hosted or customer-hosted. It is a different product from Google's free Data Studio (formerly Looker Studio).

Metabase is a browser-based BI tool from Metabase, Inc. The open source edition is AGPL licensed; paid features use the Metabase Commercial Software License. The latest release on GitHub when checked was 63.19 (1 October 2026). Business users build questions in a graphical query builder, analysts write SQL in the native query editor, and curated models and metrics give both a shared starting point. You self-host it as a JAR or Docker container, or buy Metabase Cloud.

Side by side

Key differences

Semantic layer: LookML in Git versus Metabase models and metrics

Looker makes the semantic layer code. Beyond views, dimensions and measures, LookML supports derived tables: either native (built from existing LookML fields) or SQL-based, written in your database's dialect. Looker's documentation gives this pattern for a SQL-based derived table, which can also be persisted to a scratch schema on a schedule:

# LookML (SQL-based derived table)
view: customer_order_summary {
  derived_table: {
    sql:
      SELECT
        customer_id,
        MIN(DATE(time)) AS first_order,
        SUM(amount) AS total_amount
      FROM orders
      GROUP BY customer_id ;;
  }
}

Every change goes through Git and Development Mode, so a metric can be reviewed like any other code before it reaches dashboards.

Metabase builds its semantic layer in the interface. A model is a curated dataset created from the query builder or from SQL, with editable metadata such as display names, descriptions and column types; Metabase notes that SQL models need that metadata filled in by hand. A metric is a saved aggregation (for example Sum([Price])) on a table, question or model, and Metabase calls metrics "the official way to calculate important numbers for your team". Metabase documents that metrics are created in the query builder and are available only there, not in the SQL editor, where snippets play a similar role. On Pro and Enterprise, Remote Sync exports dashboards, questions, models and library content as YAML files to a Git repository (GitHub, GitLab or Bitbucket), so a read-only production instance can be deployed from Git; Metabase notes that all users of one instance work on a single branch at a time. In our assessment, Looker gives stronger guarantees that a number means the same thing everywhere; Metabase gives a faster, less formal route to the same idea.

Writing SQL in each tool

Looker's ad hoc SQL tool is SQL Runner. It runs queries, browses schemas, keeps query history, and can turn a query into an ad hoc Explore or a derived table for a LookML project. It needs the See LookML and Use SQL Runner permissions, so in practice it belongs to Developer users, not Viewers. Business users explore through LookML-defined Explores and never see the SQL unless they open the SQL tab.

Metabase's native query editor is open to anyone with native query permissions on a database. It adds filter widgets through variables and field filters, reuses code through snippets, and treats models and saved questions as tables. This query follows Metabase's documented syntax:

-- Metabase native query (PostgreSQL), reading from a model
SELECT region, count(*) AS customers
FROM {{#1-customer-model}} AS c
GROUP BY region
ORDER BY customers DESC;

Metabase documents that the editor runs single read statements only: no multiple statements, stored procedures or DDL. In our view, Metabase is the more direct tool for an analyst who just wants to write a query and share the result; Looker is designed so that most users never need to.

Databases and where queries run

Both tools query the database live rather than storing a copy, so query cost and speed depend on the database underneath. Looker lists more than 40 supported SQL dialects, including BigQuery, Snowflake, Redshift, PostgreSQL, MySQL, SQL Server, Databricks and Oracle. Metabase ships official drivers for Athena, BigQuery, ClickHouse, Databricks, Druid, MongoDB, MariaDB, MySQL, Oracle, PostgreSQL, Presto, Redshift, Snowflake, SparkSQL, SQL Server, SQLite, Starburst and Vertica, with community drivers for others. MongoDB is notable: Metabase supports it officially, and Looker is limited to SQL dialects. On a pay-per-query warehouse, both tools' live queries are billed by the warehouse.

Hosting and licences

Looker (Google Cloud core) is a Google-managed service on Google Cloud, with Enterprise edition features such as private connections, customer-managed encryption keys and VPC Service Controls. Customer-hosted Looker (original) is the option for running it on your own infrastructure. There is no open source edition.

Metabase is a single Java service: the docs recommend Eclipse Temurin JRE 25 for the JAR, or Docker, with PostgreSQL or MySQL as the application database in production. The AGPL open source edition is free with unlimited users; SSO, row and column-level permissions and SSO-based embedding need a paid plan, self-hosted or on Metabase Cloud. The AGPL requires you to offer modified source if you modify Metabase and make it available to others over a network.

Embedding and cost model

Looker offers private embedding in every edition; signed and cookieless embedding for customer-facing apps needs the Embed edition, which also has the highest API limits. Looker's BI connectors also let Power BI, Tableau Desktop, Excel, Google Sheets and Data Studio read LookML models. Metabase allows public links, guest (signed) embeds and full-app embedding on every plan including open source, while modular embedding with SSO, the Embedded analytics SDK features and row-level data segregation need Pro or Enterprise.

The cost models differ in kind. Looker is a negotiated annual commitment for a platform edition plus user licences. Metabase publishes its plan prices, so you can calculate the cost of Starter or Pro yourself, and the open source edition costs only the hosting.

Pricing and licensing

Looker. Google's Looker pricing page, checked 7 October 2026, describes a platform fee by edition (Standard, for organisations with fewer than 50 users; Enterprise; Embed) plus user licences by type (Developer, Standard, Viewer), with annual commitments. Each edition includes 10 Standard users and 2 Developer users. Google publishes no list prices; you contact sales. Trial instances of Looker (Google Cloud core) last 90 days. Database or warehouse charges for the queries Looker runs are separate.

Metabase. The open source edition is free. Listed on Metabase's pricing page in October 2026 (USD): Starter USD 100 per month with 5 users included, then USD 6 per user per month; Pro USD 575 per month with 10 users included, then USD 12 per user per month; Enterprise custom, from USD 20,000 per year. Yearly billing is discounted, Starter and Pro have a 14-day trial, and paid plans are available as cloud or self-hosted.

Pricing checked on the vendors' official pages on 7 October 2026. Prices change; confirm before buying.

Where each one leads

Limitations

When to choose each

When neither is right

• You want an open source tool with a full SQL IDE and an Apache 2.0 licence: see Metabase vs Apache Superset.
• You are a Microsoft organisation that wants a semantic model with published per-user prices: see Power BI vs Metabase and Power BI vs Looker.
• Visual, drag-and-drop exploration is the priority: see Tableau vs Metabase and Power BI vs Tableau.
• You only need simple reports on Google data: Data Studio (formerly Looker Studio) is free; see Looker vs Looker Studio.

Final recommendation

Frequently asked questions

Sources

• Google Cloud: Looker pricing
• Looker docs: Looker (Google Cloud core) overview
• Looker docs: Editions and entitlements
• Looker docs: Supported dialects
• Looker docs: Derived tables
• Looker docs: SQL Runner basics
• Looker docs: Git version control
• Looker docs: BI connectors
• Metabase pricing
• Metabase: models
• Metabase: metrics
• Metabase: Remote Sync
• Metabase: SQL editor
• Metabase: connecting to databases
• Metabase: embedding overview
• Metabase on GitHub (licences and releases)

Checked October 2026.

How we research comparisons: our editorial method.