Apixly · API pricing
Qwen3-Reranker-0.6B
Qwen3-Reranker-0.6B · Reranking. Multilingual candidate scoring; accepts a task instruction and up to 200 candidate documents per request.
MultilingualInstruction supportUp to 200 candidates
Gateway price for Qwen3-Reranker-0.6B ↓API integration · Reranking
Use reranking after retrieving candidate documents. Submit a query and candidate documents, then use the returned relevance scores and original document indices to reorder the candidates.
- Request model ID
- Qwen3-Reranker-0.6B
- Public groups
- Embedding / Rerank
Create an Apixly API key, bind a group that lists this exact model, and keep the key in a server-side environment variable. Public prices exclude personal discounts; check the selected group before sending a request.
curl --fail-with-body --max-time 120 'https://apixly.ai/v1/rerank' \
-H "Authorization: Bearer $APIXLY_API_KEY" \
-H 'Content-Type: application/json' \
--data-binary '{
"model": "Qwen3-Reranker-0.6B",
"query": "fruit",
"documents": [
"Apples and pears",
"A train timetable"
]
}'Start with one small request. Check the complete result and reported usage, then match the request to its charge in the console. For chat, inspect choices[0].message.content; embeddings return data and reranking returns results. A successful HTTP status alone is insufficient. This example makes no automatic retry.
Compare related models
Public quotes
Public prices
Providers
- 1–25 inputs per request/ 1,000 input1–25 inputs per request
- $0.036
1–25 inputs per request/ 1,000 input
Channel: Embedding / Rerank
26+ inputs per request/ 1M tokens
- Input/ 1M tokens26+ inputs per request
- $0.04
- Output/ 1M tokens26+ inputs per request
- $0.04
Official prices and discounts use the same billing unit and tier.
Each price shows its own billing unit. Input, output, cache and reference media are charged separately where listed.