TL;DR
Thinking Machines, led by Mira Murati, announced development of ‘interaction models’ that allow AI to respond in real time across audio, video, and text. A limited research preview is expected soon, with wider release later this year.
Thinking Machines, the AI startup founded by former OpenAI CTO Mira Murati, announced on May 11, 2026, that it is developing ‘interaction models’ designed to enable AI systems to respond in real time across multiple modalities, including audio, video, and text.
The company explained that current AI models process information in a single thread, waiting for users to finish speaking or typing before generating responses, which limits the depth of human-AI interaction. Their new ‘interaction models’ aim to overcome this bottleneck by allowing AI to continuously perceive and respond as users communicate naturally, in real time.
Examples provided by Thinking Machines include AI listening for mentions of animals during a story, translating speech instantly, and even detecting when a user is slouching. These capabilities suggest a more seamless and intuitive human-AI collaboration experience.
Mira Murati founded Thinking Machines in February 2025 after leaving OpenAI. The company has experienced some turnover, with key members leaving for Meta and OpenAI, but continues to pursue advanced AI research.
Why It Matters
This development could significantly impact how humans interact with AI, making these systems more responsive, intuitive, and capable of understanding complex, real-time inputs across multiple modalities. Such advancements could influence applications in communication, accessibility, and productivity tools, shaping the future of AI-human collaboration.
real-time multi-modal AI interaction device
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Background
Murati’s departure from OpenAI in early 2025 marked a notable shift, and her new company, Thinking Machines, has been focused on pushing the boundaries of AI interaction. The concept of real-time, multi-modal AI responsiveness addresses long-standing limitations in current models, which typically process one input at a time without continuous perception.
While the company has shared examples and a vision for more natural human-AI interactions, it has not yet released its interaction models publicly. A limited research preview is expected in the coming months, with a broader rollout later this year.
“We believe we can solve the bandwidth bottleneck by making AI interactive in real time across any modality.”
— Mira Murati, Founder of Thinking Machines
“Our models enable AI to continuously take in audio, video, and text, think, respond, and act in real time, meeting humans where they are.”
— Thinking Machines spokesperson
Language Translator Device No WiFi Needed, 2026 Upgraded AI Translator, Support 150 Languages Voice Instant Two-Way Translation, Offline/Photo Translator for Business Travel
【AI Translator Supporting 150 Languages】A20 AI translator adopts the latest technology, ultra-fast and accurate translation, the response time…
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What Remains Unclear
Details about the technical capabilities of the interaction models, their performance benchmarks, and the scope of the upcoming research preview remain unclear. It is also uncertain when a wider commercial release will occur and how these models will be integrated into existing AI platforms.
details of ai tools: ai
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
What’s Next
Thinking Machines plans to launch a limited research preview in the coming months, allowing select users to test the interaction models. The company also intends to expand the models’ capabilities and gather feedback before a broader release later this year.
The Human-Agent Orchestrator: Leading and Scaling AI-Driven Organizations
As an affiliate, we earn on qualifying purchases.
As an affiliate, we earn on qualifying purchases.
Key Questions
What are interaction models?
Interaction models are AI systems designed to process and respond to multiple types of inputs—such as audio, video, and text—in real time, enabling more natural human-AI collaboration.
When will the models be available to the public?
Thinking Machines plans to offer a limited research preview in the coming months, with a wider release expected later this year.
How do these models differ from existing AI systems?
Current AI models typically process one input at a time and generate responses after the input is complete. The new interaction models aim to operate continuously and respond instantly across multiple modalities, mimicking natural human conversation.
Who is Mira Murati, and what is her background?
Mira Murati is the founder of Thinking Machines and former CTO of OpenAI, known for her leadership in developing advanced AI technologies.
What are the potential applications of these interaction models?
Potential applications include more intuitive virtual assistants, real-time translation, accessibility tools, and enhanced communication interfaces that respond seamlessly to multi-modal inputs.
