Cohere releases Rerank 5 with passage-level retrieval and calibrated scores
In short: Cohere launched Rerank 5, its newest enterprise reranking models, in Pro and Fast tiers. Rerank 5 Pro leads major public benchmarks including BRIGHT, ViDoRe V3, BEIR and nanoMTEB, while Rerank 5 Fast offers near-leading quality at lower cost and higher throughput. The release introduces Context Focus, which returns the most relevant text spans within documents rather than whole documents, and calibrated relevance scores that read as probabilities across models and domains.
This summary was generated automatically by AI from Cohere's publication. It is our own text, not a copy of the original — facts, figures and quotes belong to the source, linked above and below.
What changed?
- 1Rerank 5 Pro and Fast launched, available via Cohere API, Model Vault, Microsoft Foundry, Amazon SageMaker, VPC/on-premises deployment
- 2Pricing: $0.05 per 1M tokens for Pro, $0.02 for Fast
- 3Rerank 5 Pro leads on BRIGHT, ViDoRe V3, BEIR and nanoMTEB benchmarks, ranking first on 39 of 46 datasets tested
- 4New Context Focus capability returns scored passages within documents in a single API call, cutting LLM token usage with minimal accuracy loss
- 5Relevance scores are now calibrated: a score of 0.7 corresponds to roughly a 70% chance the document is useful
- 6Multilingual gains across 19 non-English languages, including the six priority business languages tested
| Parameter | Before | Now |
|---|---|---|
| Context Focus (passage-level retrieval) | Not specified | New feature: single Rerank call returns scored text spans within documents, not just document-level scores |
| BRIGHT benchmark score (reasoning-intensive retrieval) | Rerank 4 Pro: 69.8 (82.2 minus 12.4 gain) | Rerank 5 Pro: 82.2 |
| Relevance score calibration | Uncalibrated scores, model-specific thresholds | Calibrated probability scores: 0.7 means ~70% chance document is useful |
| Throughput (Fast vs Pro) | Not specified | Rerank 5 Fast: ~187 docs/sec on A100, 2.9x throughput of Pro, ~3x faster response times |
| Pricing | Not specified | $0.05 per 1M tokens (Pro), $0.02 per 1M tokens (Fast) |
Why it matters
For teams building retrieval-augmented agents, reranking quality and token efficiency directly affect LLM cost, latency and answer accuracy; calibrated scores also make it easier to set a single confidence threshold instead of tuning cutoffs per model or domain.
What it means for AI agents and contact centers
Teams building voice agents or contact-center tools that query knowledge bases or CRM documents before answering could use Rerank 5's Context Focus to send only the most relevant passages to the LLM, reducing per-call token cost and latency — worth comparing against the current retrieval pipeline, especially for multilingual or document-heavy use cases.
🧪 Worth Testing
Context Focus and calibrated scores offer a concrete way to cut token costs and set one relevance threshold across languages, which is directly testable against an existing retrieval setup.
Sources
- CohereOfficialPrimary sourceOriginal article →„Introducing Rerank 5: precise, efficient enterprise document retrieval Give your AI the most relevant context for each request, now with passage-level precision. Oct 21, 2026 8 min read“21 Oct 2026, 18:00
- Published by source
- 21 Oct 2026, 18:00
- Found by our system
- 21 Oct 2026, 17:48
- Summary generated
- 21 Oct 2026, 17:50
This article was written by AI from the original source. Facts, numbers and prices come from the source; missing values are marked “Not specified”. Legal notice, copyright and privacy