Every cell below comes from a real run against a shared DuckDB (and Snowflake trial) dataset. The same Jaffle-shop questions run through Semantic Rails, MetricFlow, Cube, Malloy, Snowflake Semantic Views, and KtX. Where a layer can't express the intent as a governed primitive, the question still gets answered - but the cost shows up as a helper view, a precomputed rollup, or query-level SQL. The Cube row is a captured run with its source, results, audit, SBOM, and original lock graph pinned for offline review.
Coming from MetricFlow? Translate your existing project with one command.
The translator reads your semantic_models/
directory or semantic_manifest.json, and
emits a complete package directory you can load, edit, and
commit. Imported metrics keep their semantics and are
easily upgraded to the more expressive primitives used by
Semantic Rails.
uv run semantic-rails import \
--from metricflow \
--source path/to/target/semantic_manifest.json \
--output configs/semantic_rails \
--package-id my_package Any agentic governance solution should be able to answer
the first 7 questions as table-stakes. The other 9
questions are still relatively basic, but demonstrate
where the different solution architectures deviate.
Count, sum, group-by-month. Every layer can basically do the basics. Otherwise they wouldn't be in the conversation.
| Semantic Rails | 7 native |
| MetricFlow | 7 native |
| Cube | 6 native · 1 workaround (q05) |
| Malloy | 7 native |
| Snowflake SV | 7 native |
| KtX | 7 native |
Metric predicates, temporal validity, same-store conversion, contextual entity-graph inheritance. Without these as primitives, your agent leaves the governed API and writes raw SQL against the warehouse.
| Semantic Rails | 9 native |
| MetricFlow | 2 native · 7 precomputed |
| Cube | 3 workaround · 6 precomputed |
| Malloy | 9 workaround |
| Snowflake SV | 9 workaround |
| KtX | 9 workaround |
Sixteen questions that any analyst should be able to answer.
All 16 normalized outputs match across the six layers; configuration parity rules are published in the methodology.
| Question | Semantic Rails | MetricFlow | Cube | Malloy | Snowflake SV | KtX |
|---|
Your semantic layer needs to work with your data. Since
we're open-source, it's easy to add more dialects and more
connectors. Don't lock yourself into a proprietary
solution.
| Backend | Semantic Rails OSS adapter | dbt Semantic Layer Cloud | Cube | Malloy | Snowflake Semantic Views | KtX primary source / ktx-sl |
|---|---|---|---|---|---|---|
| Snowflake | native connector | native platform | native connector | native connector | warehouse-native | native connector |
| BigQuery | native connector | native platform | native connector | native connector | no | native connector |
| Databricks | native connector | native platform | native connector | native connector | no | dialect-only |
| Amazon Redshift | no | native platform | native connector | no | no | no |
| SQL Server | no | no | native connector | bridge | no | native connector |
| Microsoft Fabric | no | no | native connector | no | no | no |
| PostgreSQL | native connector | native platform | native connector | native connector | no | native connector |
| MySQL | no | no | native connector | native connector | no | native connector |
| ClickHouse | native connector | no | native connector | no | no | native connector |
| Amazon Athena | native connector | no | native connector | no | no | no |
| Trino / Presto | no | Trino only | native connector | native connector | no | no |
| DuckDB | native connector | no | native connector | native connector | no | dialect-only |
| MotherDuck | native connector | no | native connector | native connector | no | no |
| DuckLake | native connector | no | no | no | no | no |
| SQLite | no | no | native connector | no | no | native connector |
Every configuration, query, and normalized output is in the repo. Cube's original npm graph is preserved under non-installable evidence filenames with an offline verifier. If you find a better configuration for any layer, submit a PR and we will gladly update the benchmark.