Nearly a year after releasing ReRank 3.5, CoHire launched the latest version of its search model, now with a larger context window to help agents find the information they need to complete their tasks.
Cohere said in a blog post that ReRank 4 has a 32K context window, representing a four-fold increase compared to 3.5.
According to the blog post, “This enables the model to handle longer documents, evaluate multiple paragraphs at once, and capture relationships across sections that would be missed in smaller windows.” “Therefore, this extended capability improves ranking accuracy for realistic document types and increases confidence in the relevance of retrieved results.”
ReRank 4 comes in two flavors: Fast and Pro. As a smaller model, Fast is best suited for use cases that require both speed and accuracy, such as e-commerce, programming, and customer service. Pro is optimized for tasks that require deep reasoning, accuracy, and analysis, such as building risk models and performing data analysis.
Enterprise search has gained greater importance this year, especially as AI agents have access to more information and context about the organization they work for. Cohere said rerankers “significantly increase the accuracy of enterprise AI search by refining initial retrieval results.” ReRank 4 addresses the subtle gap created by some bi-encoder embeddings — models that help simplify retrieval augmented generation (RAG) tasks — by using a cross-encoder architecture “that jointly processes queries and candidates, capturing subtle semantic relationships and reordering the results to surface the most relevant items,” Cohere said.
Performance and Benchmarks
Cohere benchmarked the models against other reranking models, such as Quen Reranker 8b, Gina Rerank v3 from Elasticsearch, and Voyage Rerank 2.5 from MongoDB, across tasks in the finance, healthcare, and manufacturing domains. ReRank 4 performed strongly, if not better than its competitors.
ReRank 3.5 stood out because of its ability to support multiple languages, and Cohere said ReRank 4 continues that trend. It understands over 100 languages, including state-of-the-art retrieval in 10 major business languages.
Agent and reranking model
The purpose of ReRank 4 is to help agent tasks understand which data is best suited for their tasks and to provide more context.
Cohere noted that the model is a key component of its agentic AI platform, North, as it “integrates seamlessly into existing AI search solutions, including hybrid, vector, and keyword-based systems, with minimal code changes.”
As more enterprises look to use agents for research and insights, as shown by the rise of Deep Research features, models that help filter out irrelevant content, such as rerankers, become more essential.
“This is particularly impactful for agentic AI, where complex, multi-step interactions can rapidly outgrow model calls and saturate context windows,” Cohere said.
The company argues that ReRank4 helps reduce token usage and the number of retries an agent needs to get things right by preventing low-quality information from reaching the LLM.
self-education
Cohere said ReRank 4 stands out not only for its robust reranking capabilities, but also for being the first reranking model that learns on its own.
Users can customize ReRank 4 for the use cases they encounter more frequently without any additional annotated data. Like foundation models like GPT-5.2, where people can tell preferences and the model remembers them, ReRank 4 users can tell the model their preferred content types and document corpora.
For example, if used with ReRank 4 Fast, the model becomes more competitive with larger models because it is more accurate and taps the specific data users want.
“Looking further, we also explored how ReRank 4’s self-learning capabilities perform on entirely new search domains,” Cohere said. “Using a health care-focused dataset that mimics a physician’s need to obtain patient-specific information – not just expertise from a given medical discipline – we found that enabling self-learning led to consistent, substantial benefits. The result: a clear and significant increase in retrieval quality for ReRank 4 Fast across the board.”
<a href