LLM ReferenceLLM Reference

QwQ 32B on OpenRouter

QwQ · Alibaba

ServerlessOpen Source

Pricing

TypePrice (per 1M)
Input tokens$0.15
Output tokens$0.58

Capabilities

VisionMultimodalReasoningFunction CallingTool UseStructured OutputsCode Execution

About QwQ 32B

QwQ-32B is the first full release in Alibaba's QwQ reasoning series. Built on a 32.5B-parameter dense transformer, it achieves significantly enhanced performance on complex tasks—mathematics, coding, and multi-step reasoning—through extended chain-of-thought thinking. Available open-weight on Hugging Face, it delivers frontier reasoning in an efficient package.

Get Started

Model Specs

Released2025-03-05
Parameters32.5B
Context128K
ArchitectureDecoder Only