BigQuery data preparation is generally available (GA). It offers AI-powered suggestions from Gemini for data cleansing, transformation, and enrichment
Share
Services
## Feature
[BigQuery data preparation](https://cloud.google.com/bigquery/docs/data-prep-introduction) is [generally available](https://cloud.google.com/products#product-launch-stages) (GA). It offers AI-powered suggestions from Gemini for data cleansing, transformation, and enrichment. BigQuery supports visual data preparation pipelines and pipeline scheduling with Dataform.
## Feature
You can now create [remote models](https://cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-create-remote-model) in BigQuery ML based on [Llama](https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/llama) and [Mistral AI](https://cloud.google.com/vertex-ai/generative-ai/docs/partner-models/mistral) models in Vertex AI.
Use the [ML.GENERATE\_TEXT function](https://cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-generate-text) with these remote models to perform generative natural language tasks for text stored in BigQuery tables. Try this feature with the [Generate text by using the ML.GENERATE\_TEXT function](https://cloud.google.com/bigquery/docs/generate-text) tutorial.
This feature is [generally available](https://cloud.google.com/products/#product-launch-stages) (GA).
## Change
An updated version of [JDBC driver for BigQuery](https://cloud.google.com/bigquery/docs/reference/odbc-jdbc-drivers#current%5Fjdbc%5Fdriver) is now available.
## Feature
[Smart-tuning](https://cloud.google.com/bigquery/docs/materialized-views-use#smart%5Ftuning) is now supported for [materialized views](https://cloud.google.com/bigquery/docs/materialized-views-intro) when they are in the same project as one of their base tables, or when they are in the project running the query. This feature is [generally available](https://cloud.google.com/products#product-launch-stages) (GA).
## Change
BigQuery ML now uses dynamic token-based batching for [embedding generation](https://cloud.google.com/bigquery/docs/reference/standard-sql/bigqueryml-syntax-generate-embedding) requests. Dynamic token-based batching puts as many rows as possible into one request. This change boosts per-request utilization and improves scalability for any [queries per minute (QPM) quota](https://cloud.google.com/bigquery/quotas#cloud%5Fai%5Fservice%5Ffunctions). Actual performance varies based on the embedding content length, with an average 10x improvement.