Cohere: Cohere Rerank 4 Fast

Cohere Rerank 4 Fast orders retrieved documents by relevance to a query, with a focus on latency and throughput. It supports multilingual text retrieval and is billed by reported search units, making it useful between retrieval and answer generation in a RAG pipeline.

Compare models

rerank-v4.0-fast

AvailableCohereRerankSync

Price: From $0.002
Modalities: Reranking

About Cohere Rerank 4 Fast

Cohere Rerank 4 Fast is the latency-oriented model in Cohere's Rerank 4 family. It orders retrieved documents by relevance to a query and sits between retrieval and answer generation in a RAG pipeline. Cohere describes Fast as a smaller model for cases where speed and accuracy both matter. It handles multilingual text and semi-structured content such as JSON, tables and code-like text.

Where it works well

Low latency and high throughput, which is how Cohere positions Fast against Pro.
Multilingual reranking, so one model can serve queries and documents in many languages.
Semi-structured input such as JSON, tables and code-like content, per Cohere's Rerank 4 announcement.
A longer document window than the previous Rerank generation, so longer passages fit.
Self-learning adaptation is a Rerank 4 feature, as Cohere describes it.

When to choose another model

Cohere builds Rerank 4 Pro for deeper reasoning and the highest precision, so hard relevance cases favour Pro.
It only scores documents; it does not generate answers or embeddings.
Quality depends on first-stage retrieval, since it can only reorder candidates it receives.

Getting started

Create API key
Create a key in Console, then use it with every model on the platform.

Send your first request

Copy the example for your language and run it against the endpoint.

RerankingTokenLab endpoint

POST/v1/rerank

curl -X POST "https://api.tokenlab.sh/v1/rerank" \
  -H "Authorization: Bearer sk-xxx" \
  -H "Content-Type: application/json" \
  -d '{
  "operation": "rerank",
  "model": "rerank-v4.0-fast",
  "query": "What is the capital of France?",
  "documents": [
    "Paris is the capital of France.",
    "Berlin is the capital of Germany."
  ]
}'

Pricing

The Verified price applies to Verified, which costs less on most models. The Official price is the model maker's published price and applies to the more reliable Official route. Auto bills the route that completes the request.

Reranking

per request

Official price: $0.002
Verified price: —
Discount: —

	Official priceper request	Verified priceper request	Discount
Reranking	$0.002	—	—

Usage & activity

Success rate is the share of requests that completed. Latency is how long a full response takes; P95 means 95% of requests finished within that time.

Usage & availability

Last 24 hours

Requests
Success rate
P95 latency
Total tokens

Model performance

Last updated Oct 14, 2026, 11:00 PM UTC

Metrics appear once privacy and data volume thresholds are met.

Data is based on aggregate user requests, excluding status checks.

Open in Console

Open Cohere Rerank 4 Fast in Console with a prompt ready to edit or send.

Help me try rerank-v4.0-fast with rerank at /v1/rerank. Show the result and cost.

Use cases

Live search ranking
Rerank the top fifty hits from keyword or vector search before showing them, where each query must respond quickly.
Chat retrieval
Pick the best passages to put in front of a language model in a support or documentation assistant.
Large candidate sets
Process many candidate documents per query in batch jobs, such as nightly relevance checks, without waiting long for each request.

Prompt examples

Query: how do I rotate an API key without downtime? Documents: five runbook sections.

Query: Kündigungsfrist Mietvertrag. Documents: ten German and English legal paragraphs.

Query: which invoice mentions late fee? Documents: JSON records of twenty invoices.

FAQ

What does Rerank 4 Fast do?

It scores a list of documents against a query and returns them ordered by relevance. Use it after a first retrieval step to improve the order before documents reach a user or a language model.

How is Fast different from Pro?

Cohere describes Fast as optimized for low latency and high throughput, and Pro as optimized for quality and complex cases. Fast suits live paths; Pro suits harder relevance judgments.

Does it work in other languages?

Yes. Cohere lists multilingual reranking for the Rerank 4 family, so queries and documents can be written in different languages, and one model can serve a mixed-language collection.

Can it rank JSON or tables?

Cohere says Rerank 4 supports semi-structured content including JSON, tables and code-like text. Format each record consistently, with the same field order, for the most stable ordering.

How is it billed?

By reported search units rather than by tokens or text length, so cost follows the number of search requests you send, not how long each document is. This differs from token-billed chat and embedding models.

How much does Cohere Rerank 4 Fast cost?

On TokenLab, Cohere Rerank 4 Fast costs $0.002 per request. The pricing table above shows the full breakdown. Rates depend on the billing unit, specification, and usage. Compare matching conditions in the model's detailed pricing; a single rate does not determine the total cost.

Which endpoint should Cohere Rerank 4 Fast use?

Use https://api.tokenlab.sh/v1/rerank for Cohere Rerank 4 Fast. The request example below shows the matching code shape.

Which operations does Cohere Rerank 4 Fast support?

Cohere Rerank 4 Fast supports Reranking. Select an operation above to see its endpoint and request example.