Jina Reranker m0
Jina Reranker m0 ranks visually rich documents, including pages with charts, figures and tables. It can score text and image content against a query without requiring the entire document to become plain text first.
How the model works
The 2.4-billion-parameter model adapts Qwen2-VL-2B-Instruct with LoRA and a trained scoring head. Pairwise and listwise losses teach relevance ranking. The combined query and document limit is 10,240 tokens. This is a vision-language scoring design, distinct from the earlier text-only cross-encoders.Jina Reranker m0 model details
Use and access
The model card documents text-to-text, text-to-image, image-to-text and mixed-content tasks. Jina offers weights, a GGUF conversion, API access and cloud marketplace deployments. The downloadable weights use CC BY-NC 4.0. A high relevance score identifies a useful candidate; it does not establish that the document is true or that extracted text is correct.Jina Reranker m0 model detailsJina Reranker m0 technical overview
Specifications & access
| Detail | Value and conditions |
|---|---|
| Launch date | |
| Access | API, Downloadable |
| Model identifiers | jina-reranker-m0jinaai/jina-reranker-m0jina-reranker-m0-GGUF |
| Context window | 10,240 tokens Combined query and document input, including visual tokens. Jina Reranker m0 model details |
| Parameters | 2.4 billionJina Reranker m0 model details |
| Weight license | CC BY-NC 4.0Jina Reranker m0 model details |
| Base model | Qwen2-VL-2B-InstructJina Reranker m0 model details |
Related models
- Jina Reranker v3.5
Jina AI · Jina Reranker
A listwise search ranker tuned for specialist documents and structured data.
Launch date API · Downloadable - Jina Embeddings v4
Jina AI · Jina Embeddings v4
A visual-document encoder with both single-vector and token-vector retrieval.
Launch date API · Downloadable