Thinking Machines Lab, an AI startup founded by former OpenAI CTO Mira Murati, has announced its latest project: interaction models designed to facilitate real-time conversation. The company aims to revolutionize AI communication, making it more akin to a phone call than a traditional text exchange.
AI Like a Phone Call
Current AI models operate in a back-and-forth manner—users speak, the AI listens, and then responds. Thinking Machines is challenging this norm with a model that processes input and generates responses simultaneously, reflecting a more dynamic and interactive experience.
Introducing TML-Interaction-Small
The technical term for this innovation is “full duplex,” and Thinking Machines' model, TML-Interaction-Small, reportedly delivers responses in 0.40 seconds. This speed rivals natural human conversation and outpaces existing models from major players like OpenAI and Google.
Research Preview and Future Releases
Currently, TML-Interaction-Small is in the research preview phase and not publicly available. Thinking Machines plans a limited research preview in the coming months, with broader accessibility anticipated later in the year. Whether the model will meet its technical benchmarks in real-world applications remains to be seen.
Potential Impact
While benchmarks are promising, the true test will be user experience once the model is accessible. The concept of integrating interactivity into AI models could redefine user expectations and set new standards in AI-driven communication.
Source: https://techcrunch.com/2026/05/11/thinking-machines-wants-to-build-an-ai-that-actually-listens-while-it-talks/
