Thinking Machines Lab details interaction models, which can think and respond in real time, letting users and AI interact continuously for better collaboration

Thinking Machines Lab just dropped a research preview of what it calls "interaction models" — AI systems designed to handle real-time, continuous interaction natively. No external scaffolding. No bolted-on workarounds. The thinking happens live, as the conversation flows.

The core idea: instead of the usual prompt-wait-respond loop, these models can think and reply simultaneously while a user is still engaging. It's a fundamentally different architecture for human-AI collaboration, treating interaction as a first-class capability rather than an afterthought layered on top.

The lab is positioning this as a shift in how AI systems are built. Traditional models treat conversation as a series of discrete exchanges. Interaction models blur that line, enabling something closer to genuine back-and-forth dialogue.

It's still a research preview, so temper expectations. But the direction is clear: AI that collaborates, not just responds.