Introducing Rerank 5: precise, efficient enterprise document retrieval

Cohere released Rerank 5, its enterprise retrieval models in Pro and Fast tiers, which it says are its most advanced yet, and priced API access at $0.05 per 1M tokens for Pro and $0.02 for Fast.

• Both tiers add Context Focus, which returns the most relevant passages within each document rather than whole documents or chunks, cutting the tokens an LLM reads and therefore its cost.
• Cohere says Rerank 5 Pro leads public benchmarks including ViDoRe V3, nanoMTEB, BRIGHT and BEIR, ranking first on 39 of 46 composite datasets and at least second on 43.
• On BRIGHT, which tests reasoning-intensive retrieval, Cohere reports Pro scores 82.2, ahead of Voyage AI Rerank 3 by 5.8 points and Rerank 4 Pro by 12.4 points.
• Rerank 5 Fast is the speed-optimized tier: Cohere says it reranks about 187 standard-length documents per second on an NVIDIA A100 GPU, 2.9x the throughput of Pro, with nearly 3x faster response times.
• The models are generally available on the Cohere API, Model Vault, Microsoft Foundry, Amazon SageMaker, privately in a VPC or on-premises, and to all North users.