Qwen3-VL Reranker 8B
JSON →A vision-language reranker from the Qwen3 family with 8 billion parameters, capable of re-ranking both text and image inputs.
Specs
context window 32K tokens
max output 32K tokens
Capabilities
visionstreaming
Dates
releasedApr 2025
Resources
API
full doc /v1/models/Qwen/Qwen3-VL-Reranker-8B