tejasunku/qwen3-reranker-8b 📝🔢 → 📝

▶️ 5 runs 📅 Nov 2025 ⚙️ Cog 0.16.9 🔗 GitHub 📄 Paper
passage-ranking retrieval text-reranking

About

Rerank query search results using qwen3 reranker 8b model, with support for instructions

Example Output

Output

{
"instruction": "Given a web search query, retrieve relevant passages that answer the query",
"query": "What is the capital of China",
"results": [
{
"document": "The capital of China is Beijing",
"score": 0.457763671875
},
{
"document": "Gravity is a force that attracts two bodies towards each other. It gives weight to physical objects and is responsible for the movement of planets around the sun",
"score": -0.8974609375
}
],
"total_documents": 2
}

Performance Metrics

1.20s Prediction Time
351.54s Total Time
All Input Parameters 
{
  "query": "What is the capital of China",
  "top_k": 5,
  "documents": "[\"The capital of China is Beijing\", \"Gravity is a force that attracts two bodies towards each other. It gives weight to physical objects and is responsible for the movement of planets around the sun\"]",
  "batch_size": 8,
  "instruction": "Given a web search query, retrieve relevant passages that answer the query"
}
Input Parameters
query (required) Type: string
Query text for reranking
top_k Type: integerDefault: 5Range: 1 - 100
Number of top documents to return
documents Type: stringDefault: ["Document 1 text", "Document 2 text", "Document 3 text"]
JSON string containing list of documents to rerank
batch_size Type: integerDefault: 8Range: 1 - 32
Batch size for processing documents
instruction Type: stringDefault: Given a web search query, retrieve relevant passages that answer the query
Task instruction for the reranker (recommended for better performance)
Output Schema

Output

Type: string

Version Details
Version ID
1fb2f6046d0f044dd6b3a13cc7ff61d33973619716991b25d346b455972cec99
Version Created
November 22, 2025
Run on Replicate →