Sold by: Cohere
Deployed on AWS
Rerank 5 Fast is Cohere's speed-optimized reranking model, built for high-volume, latency-sensitive search and agents. It outperforms our previous flagship, Rerank 4 Pro, and returns the passages that matter within each document, so your LLM gets the right context with fewer tokens.
Overview
Every token your LLM reads is a token you pay for, and retrieval decides which tokens those are. Rerank 5 Fast reads each query alongside the results from your existing keyword, vector or hybrid search and puts the most relevant first, so your AI applications get the right context for every request. Like Rerank 5 Pro, it includes Context Focus, which also returns the passages within each document that answer the query, in the same call. That means more accurate answers, fewer wasted LLM tokens, and better end-user experiences.
Rerank 5 Fast is for the live request path, wherever the reranker runs many times per request or under a tight latency budget. Smaller than Pro, it outperforms our previous flagship, Rerank 4 Pro, and much larger open models, so it's a strong default rather than a compromise. Keep conversational shopping agents responsive, pick the right tool from thousands at every agent step, suggest help articles as users write support tickets, or answer advisor questions during live client calls.
Choose Rerank 5 Pro where ranking quality decides the outcome, such as complex queries over long, domain-specific documents. Both models share the same request and response format, so you can route traffic between them as your needs change.
Key specifications:
- 32K-token context window for long documents
- Context Focus returns scored relevant passages in the same call (opt-in)
- Calibrated relevance scores, consistent with Rerank 5 Pro
- Supports 100+ languages, including cross-language search
- Reranks text and semi-structured data such as JSON, tables and code
- Works after any keyword, vector or hybrid search, with no re-indexing
- Runs in your own AWS account, so your data stays in your environment
Highlights
- Speed without compromise. Rerank 5 Fast outperforms our previous flagship, Rerank 4 Pro, on major public benchmarks at a fraction of the size. It also beats other lightweight rerankers, such as Voyage AI Rerank 3 Lite, and much larger open models. Like Pro, it performs strongly on multilingual content and parsed documents, making it a reliable default for production search.
- Best for volume. Rerank 5 Fast is optimized for high-throughput, low-latency reranking. In our tests, it reranked close to three times as many documents per second as Rerank 5 Pro, with much faster response times. That makes it the right choice for customer-facing search, support assistants, and agents that run many searches per task.
- Retrieve alongside Rerank 5 Pro. Rerank 5 Fast shares Rerank 5 Pro's request and response format, Context Focus and calibrated relevance scores, so one relevance threshold works across both. Route latency-sensitive traffic and agent loops to Fast and quality-critical queries to Pro, or shift workloads between them over time. Balance quality, speed and cost on one consistent retrieval setup.
Details
Sold by
Categories
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.g5.2xlarge Inference (Batch) Recommended | Model inference on the ml.g5.2xlarge instance type, batch mode | $3.21 |
ml.g5.xlarge Inference (Real-Time) Recommended | Model inference on the ml.g5.xlarge instance type, real-time mode | $3.00 |
ml.g5.2xlarge Inference (Real-Time) | Model inference on the ml.g5.2xlarge instance type, real-time mode | $3.21 |
ml.p5.4xlarge Inference (Real-Time) | Model inference on the ml.p5.4xlarge instance type, real-time mode | $3.12 |
ml.p4d.24xlarge Inference (Real-Time) | Model inference on the ml.p4d.24xlarge instance type, real-time mode | $4.55 |
Vendor refund policy
No refunds available.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
Rerank 5 Fast is Cohere's speed-optimized reranking model for high-volume, latency-sensitive search and agents, and it outperforms Rerank 4 Pro. It reads each query alongside the results from your existing search and returns them ordered by relevance, with calibrated scores. Like Rerank 5 Pro, it supports Context Focus (opt-in), which returns the most relevant passages within each document in the same call. Supports 100+ languages, semi-structured data and a 32K-token context window. Uses the same request and response format as Rerank 5 Pro.
Additional details
Inputs
- Summary
The model accepts JSON requests that specify a query and the list of documents to be reranked. The maximum number of documents that can be passed into a single rerank call is 1,000. The documentation below is for Version 2 of the Rerank API.
{"model": "...", "query": "...", "documents": ["..."], "top_n": 10}
Context Focus (opt-in, off by default): add relevant_spans, for example "relevant_spans": {"mode": "enabled", "threshold": 0.3, "max_spans": 10}.
- Limitations for input type
- Text input only. Real-time inference only; batch transform is not supported. Send at most 1,000 documents per request. The query and each document share a 32,768-token context window. Documents that exceed it are split into chunks and scored per chunk, and the highest chunk score is returned. Format structured data such as JSON or tables as YAML strings and keep key order consistent (in Python, yaml.dump(doc, sort_keys=False)), because long documents are truncated. Rerank 5 does not support instructions that change how relevance is judged. Context Focus works best for targeted queries: broad requests such as summarization need the full document, and tables, forms and code split less cleanly than prose. TBC: confirm whether max_tokens_per_doc is supported for Rerank 5 on SageMaker (and its default and maximum), and the query token limit.
- Input MIME type
- application/json
{"model": "rerank-v5.0-fast", "query": "What is the capital of the United States?", "documents": ["Carson City is the capital city of the American state of Nevada.", "Washington, D.C. is the capital of the United States.", "Capital punishment has existed in the United States since before it was a country."], "top_n": 2}
With Context Focus: the same request with "relevant_spans": {"mode": "enabled", "threshold": 0.3, "max_spans": 10} added.
Batch transform unavailable
Input data descriptions
The following table describes supported input data fields for real-time inference and batch transform.
Field name | Description | Constraints | Required |
|---|---|---|---|
query | The search query. The query and document share the model's 32,768-token context window. | Type: FreeText | Yes |
documents | A list of texts to compare to the query. For optimal performance, send no more than 1,000 documents per request. Format structured data (such as JSON or tables) as YAML strings for best performance. | Type: FreeText
Maximum: 1,000 documents | No |
top_n | Limits the number of returned rerank results to the specified value. If not passed, all rerank results are returned. | Type: Integer Minimum: 1 | No |
relevant_spans | Optional object that enables Context Focus, which returns the most relevant passages within each document. Fields: mode (enabled or disabled), threshold (a higher value removes more text; example 0.3) and max_spans (maximum spans per document; if not set, there is no limit; example 10). Off by default. | Type: Object. Fields: mode (enabled | disabled), threshold (number), max_spans (integer). TBC: the default threshold is currently 0.2 and is not final. | No |
Resources
Vendor resources
Support
Vendor support
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
Rerank improves search systems by sorting documents based on their semantic similarity to a query. Choose fast, when you need a fast Reranking model. Choose pro, when you need the Reranking model with the best performance.
Rerank improves search systems by sorting documents based on their semantic similarity to a query. Choose fast, when you need a fast Reranking model. Choose pro, when you need the Reranking model with the best performance.
Rerank improves search systems by sorting documents based on their semantic similarity to a query.
Embed 5 is Cohere's most powerful embedding family yet, delivering frontier retrieval on complex enterprise data that matches or exceeds the competition and makes major gains on Embed 4.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.