Alibaba Slashes Voice AI Prices by 95%

🔥 What happened Alibaba dropped Qwen-Audio-3.1, a 5-model audio stack covering ASR, TTS, and realtime interaction. The flagship Qwen-Audio-3.1-Realtime is a full-duplex speech model for voice agents that call tools. Prices got slashed: up to 95% cheaper than before. 💡 Why it matters At $6.4 per 1M audio input tokens, it undercuts OpenAI's $32 and Google's $3.00. With 262K context and features like web search and structured outputs, voice AI just became affordable for indie devs. But no open weights – managed API only. ⚡ Our take Alibaba is playing the price game hard – but the 1.1s interruption latency (vs. 0.38s for GPT-Realtime-2) shows corners were cut. Cheap comes at a cost: your users' patience.