Alibaba launched Qwen-Audio-3.0-Realtime, a real-time speech interaction dialogue model. The release upgrades four core areas simultaneously—stronger reasoning, Agent tool invocation, empathetic dialogue and full‑duplex interaction fluency—aiming for faster, smarter voice interaction. Two variants are offered: Plus for stronger inference and Flash optimized for speed. Target use cases include smart customer service, education and training, interactive entertainment and emotional companionship.