models / reranker
jina-reranker-v2-base-multilingual
jinaai/jina-reranker-v2-base-multilingual
A strong multilingual reranker: score query-document relevance across languages, not just English. Benchmarked against the field and hosted as a production API.
price
$0.008
/ 1k docs
API providers
0
downloads / mo
1.8M
throughput
512 docs/s
license
cc-by-nc-4.0
about this model
Trained by Jina AI.
The jina-reranker-v2-base-multilingual is a cross-encoder model fine-tuned for text reranking in information retrieval systems. It takes a query and a document pair and outputs a relevance score, enabling high-accuracy reranking across multiple languages.
Key Strengths
- Handles long texts up to 1024 tokens; uses a sliding window approach for inputs exceeding this length.
- Equipped with flash attention for faster inference (3x–6x speedup).
- Demonstrates competitiveness on benchmarks for text retrieval, multilingual retrieval, function-calling-aware reranking, text-to-SQL-aware reranking, and code retrieval.
Best For
- Improving search result relevance in multilingual retrieval systems.
- Reranking documents in pipelines that require high precision across diverse languages and domains (e.g., code, SQL, function calls).
License & Access
This model is licensed under CC-BY-NC-4.0 for research and evaluation. For commercial use, refer to Jina AI’s APIs or contact their sales team.
As hosted on gigarouter, the model is available as a managed, OpenAI-compatible API — no infrastructure setup required.
call it
# rerank documents by relevance; billed per document curl https://gigarouter.ai/v1/rerank \ -H "Authorization: Bearer $GR_KEY" \ -d '{"model":"jinaai/jina-reranker-v2-base-multilingual","query":"capital of France", "documents":["Paris is the capital of France.","Bananas are yellow."]}'