Cohere Rerank 3.5
Reranking model designed for production RAG applications. Improves search relevance by reordering retrieved documents based on semantic similarity to queries. Supports multilingual capabilities and handles diverse domains.
Model Information
- Provider
- Cohere
- License
- Proprietary
- Model Name
- cohere-rerank-v3.5
- Total Evaluations
- 375
Performance Record
Wins262 (34.9%)
Losses485 (64.7%)
Ties3 (0.4%)
Wins
Losses
Ties
Performance Overview
ELO ratings by dataset
Cohere Rerank 3.5's ELO performance varies across different benchmark datasets, showing its strengths in specific domains.
Cohere Rerank 3.5 - ELO by Dataset
Detailed Metrics
Dataset breakdown
Performance metrics across different benchmark datasets, including accuracy and latency percentiles.
BEIR/scifact
ELO 147445.6% WR114W-135L-1T
Accuracy Metrics
- nDCG@5
- 0.856
- nDCG@10
- 0.866
- Recall@5
- 0.891
- Recall@10
- 0.920
Latency Distribution
- Mean
- 638ms
- P50 (Median)
- 612ms
- P90
- 819ms
BEIR/fiqa
ELO 138935.2% WR88W-160L-2T
Accuracy Metrics
- nDCG@5
- 0.121
- nDCG@10
- 0.127
- Recall@5
- 0.118
- Recall@10
- 0.130
Latency Distribution
- Mean
- 510ms
- P50 (Median)
- 569ms
- P90
- 617ms
PG
ELO 133124.0% WR60W-190L-0T
Latency Distribution
- Mean
- 661ms
- P50 (Median)
- 613ms
- P90
- 792ms
Compare Models
See how it stacks up
Compare Cohere Rerank 3.5 with other top rerankers to understand the differences in performance, accuracy, and latency.