ES|QL RERANK command

The RERANK command uses an inference model to compute a new relevance score for an initial set of documents, directly within your ES|QL queries.

PUT _cluster/settings
{
  "persistent": {
    "esql.command.rerank.limit": 5000
  }
}
		

PUT _cluster/settings
{
  "persistent": {
    "esql.command.rerank.enabled": false
  }
}
		

FROM books
| WHERE title:"search query"
| SORT _score DESC
| LIMIT 100
| RERANK "search query" ON title WITH { "inference_id" : "my_rerank_endpoint" }
		

RERANK [column =] query ON field [, field, ...] [WITH { "inference_id" : "my_inference_endpoint" [, "timeout" : "<timeout_duration>"] }]
		

RERANK [column =] query ON field [, field, ...] [WITH { "inference_id" : "my_inference_endpoint" }]
		

column

(Optional) The name of the output column containing the reranked scores. If not specified, the results will be stored in a column named _score. If the specified column already exists, it will be overwritten with the new results.

query

The query text used to rerank the documents. This is typically the same query used in the initial search.

field

One or more fields to use for reranking. These fields should contain the text that the reranking model will evaluate.

inference_id

(Optional) The ID of the inference endpoint to use for the task. The inference endpoint must be configured with the rerank task type. If not specified, defaults to the preconfigured .rerank-v1-elasticsearch endpoint.

timeout

(Optional) Timeout for the inference request (for example, "30s", "1m"). If not specified, the default search timeout applies. Use this to set a per-call timeout independent of the cluster-wide search timeout.

Use RERANK to re-score search results using a machine learning model for improved relevance.

Typically, you first use a WHERE clause with a function like MATCH to retrieve an initial set of documents. This set is often sorted by _score and reduced to the top results (for example, 100) using LIMIT. The RERANK command then processes this smaller, refined subset, which is a good balance between performance and accuracy.

When using RERANK with a multivalue column, each value is ranked individually. The score column is then assigned the maximum score resulting from ranking the individual values.

The RERANK command requires an inference endpoint configured with the rerank task type. If you omit the inference_id option, RERANK uses the preconfigured .rerank-v1-elasticsearch endpoint, which is available by default.

For improved relevance, you can use the preconfigured .jina-reranker-v3 endpoint, powered by the Elastic Inference Service (EIS). To use a different model, create a rerank inference endpoint and specify its ID in the WITH clause. Refer to semantic reranking for a full list of supported reranking models.

RERANK commands may time out when processing large datasets or complex queries. The default timeout is 30 seconds.

You can set per-call timeout using the "timeout" option in the WITH clause:

RERANK "search query" ON title WITH { "inference_id": "my_inference_endpoint", "timeout": "1m" }
		

If you can't modify your timeout limits, try the following:

• Reduce data volume with LIMIT or more selective filters before the RERANK command
• Split complex operations into multiple simpler queries
• Configure your HTTP client's response timeout (Refer to HTTP client configuration)

FROM books METADATA _score
| WHERE MATCH(description, "hobbit")
| SORT _score DESC
| LIMIT 100
| RERANK "hobbit" ON description WITH { "inference_id" : "test_reranker" }
| LIMIT 3
| KEEP title, _score
		

FROM books METADATA _score
| WHERE MATCH(description, "hobbit") OR MATCH(author, "Tolkien")
| SORT _score DESC
| LIMIT 100
| RERANK rerank_score = "hobbit" ON description, author WITH { "inference_id" : "test_reranker" }
| SORT rerank_score DESC, book_no DESC
| LIMIT 3
		

FROM books METADATA _score
| WHERE MATCH(description, "hobbit") OR MATCH(author, "Tolkien")
| SORT _score DESC
| LIMIT 100
| RERANK rerank_score = "hobbit" ON title, description WITH { "inference_id" : "test_reranker" }
| EVAL original_score = _score, _score = rerank_score + original_score
| SORT _score
| LIMIT 3
		

FROM books
| WHERE MATCH(title, "return")
| RERANK "Tolkien" ON TOP_SNIPPETS(description, "Tolkien", { "num_snippets": 3, "num_words": 25 }) WITH { "inference_id" : "test_reranker" }