Ali
Qwen3-Reranker-8B
ali/qwen3-reranker-8b/v1/rerank创建于 2026-06-10 19:20:02 Qwen3-Reranker-8B是阿里巴巴通义实验室于2025年6月开源的80亿参数重排序模型，支持119种语言和32K超长文本处理。在中文检索（CMTEB-R 77.45分）、多语言检索（MTEB-R 69.02分）和代码检索（MTEB-Code 81.22分）任务中均刷新行业纪录，性能超越BGE等主流模型30%以上。 其创新单塔交叉编码器架构和指令感知能力，可显著提升检索增强生成（RAG）系统的精度与效率。
- Input / output modalities
- 文本 to 文本
- Reference input / output price
- Input¥0.5Output¥0per 1M tokens
- Context window
- 256K
- Added to catalog
- Aug 4, 2026
Providers and pricing
| Provider | Input /M | Output /M | Cached /M | Context | Max output | Details |
|---|---|---|---|---|---|---|
| Other-A-Ali | ¥0.5 | ¥0 | ¥0 | 256K | 64K |
Qwen3-Reranker-8B code examples and API guide
The upstream response does not declare a public API protocol that can be represented safely. Unknown protocols are not guessed.
Public model applications
Want to feature your application here? Click here to discuss ecosystem collaboration.No public applications are currently listed for this model.