Thinking Machines unveils AI that can interrupt users in real-time

2026-05-12 · Source: Dataconomy · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Novice, short

Summary

Thinking Machines Lab, founded by former OpenAI CTO Mira Murati, has unveiled a new AI technology called "interaction models," designed to enable real-time, interruptible conversations with AI. The company's model, TML-Interaction-Small, reportedly achieves a response time of 0.40 seconds, closely mimicking natural human conversation speed and outperforming similar models from OpenAI and Google. This "full duplex" capability allows for simultaneous input processing and response generation, departing from traditional sequential AI interactions. TML-Interaction-Small is currently in a limited research preview phase, with a wider release anticipated later this year, though its practical effectiveness awaits broader user access.

Key takeaway

For AI Architects and NLP Engineers designing conversational systems, consider adopting full duplex interaction models to enhance user experience. Thinking Machines Lab's TML-Interaction-Small demonstrates that sub-half-second response times are achievable, enabling more natural and interruptible dialogues. Evaluate this approach to move beyond sequential listen-then-respond paradigms, potentially improving engagement and utility in real-time applications like customer service or interactive assistants.

Key insights

Thinking Machines Lab developed a full-duplex AI model enabling real-time, interruptible human-AI conversation.

Principles

• AI interaction can be simultaneous, not just sequential.
• Human-like response times enhance conversational AI.
• Full duplex communication improves AI interactivity.

Method

The TML-Interaction-Small model processes input and generates responses concurrently, allowing for real-time interruption and a 0.40-second response time, mimicking natural human conversation flow.

In practice

• Implement full duplex audio/video for real-time AI interaction.
• Integrate web search and artifact generation into conversational AI.
• Translate languages in real-time during AI-human dialogues.

Topics

• Thinking Machines Lab
• Interaction Models
• TML-Interaction-Small
• Full Duplex AI
• Real-time Interaction

Best for: AI Architect, NLP Engineer, Director of AI/ML, AI Product Manager, Tech Journalist

Related on AIssential

• New Trends in AI Interaction Models — Thinking Machines Lab's Interaction Models enable full-duplex AI communication, allowing simultaneous listening and talking to eliminate conversational latency.
• Thinking Machines drops a new, highly responsive model designed for humanlike interactions in real time — New AI architecture enables full-duplex, real-time multimodal interaction by processing data in 200-millisecond micro-turns.
• Thinking Machines Just Announced More Human Like AI — Thinking Machines Lab introduces "Interaction Models" for real-time, human-like AI collaboration via simultaneous input/output.
• [AINews] Thinking Machines' Native Interaction Models - TML-Interaction-Small 276B-A12B - advances SOTA Realtime Voice and kills standard VAD — Real-time, full-duplex multimodal interaction models are advancing human-AI collaboration beyond turn-based systems.
• Thinking Machines shows off preview of near-realtime AI voice and video conversation with new 'interaction models' — Thinking Machines' interaction models enable real-time, full-duplex human-AI communication by processing multimodal inputs and outputs simultaneously.
• New AI Model SHOCKED Me — Thinking Machines Labs' new model offers real-time, multimodal, and context-aware conversational AI.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Dataconomy.