- input
- $0.00
- output
- $0.00
- Context
- 33Ktokens
- Listed
- 1w ago
Cost/1M tokens
Rerank 5 Pro is the Pro tier of Rerank 5, Cohere's latest reranking model, which the company describes as its most advanced enterprise retrieval models yet. Cohere's announcement leads with passage-level precision for enterprise retrieval, saying that with the new Context Focus feature, a single rerank call returns the most relevant passages in each document, which it claims cuts LLM token costs with minimal answer-quality loss. The model is built for reranking results after existing keyword, vector or hybrid search, giving RAG systems and agents the most relevant context per request.
How it compares
Frequently asked questions
API and pricing
Yes, Cohere lists a free Rerank 5 Pro endpoint with $0 input and output.
Yes, it has been reliable in the last day. 1 endpoint had uptime data and 1 was at 99% uptime or better, with the best uptime being 100.0%, by Cohere.
Capabilities and specs
The context window of Rerank 5 Pro is 33K tokens (32,768), and its maximum output is 29K tokens (29,491).
No, Rerank 5 Pro accepts only text as an input modality.
Quality and benchmarks
The closest alternatives, matched on model kind and release date, are rerank-3 and rerank-3-lite by VoyageAI by MongoDB, plus Nemotron 3.5 Rerank 8B by NVIDIA.
rerank-3Rerank 5 Pro vs rerank-3rerank-3-liteRerank 5 Pro vs rerank-3-liteNemotron 3.5 Rerank 8BRerank 5 Pro vs Nemotron 3.5 Rerank 8B
Versions and lineage
The listed specs show no difference. Both have a 33K token (32,768) context, both cost $0.00 input and $0.00 output per 1M tokens, and both were released on November 3, 2026.
Basics
Rerank 5 Pro was made by Cohere and was first listed on OpenRouter on November 3, 2026.
Yes, Cohere offers free API endpoints for Rerank 5 Pro. Its openness is listed as unknown.
Generated from Vertical Tabs data, updated November 9, 2026.
Discussion
No comments yet
Be the first to share your thoughts.