About us

Frontier AI Review examines how AI models are built, improved and tested—and what those advances change for the industry.

What we cover

Our five desks cover Models, Data, Training, Inference, and Evaluations. We follow model releases, architectures, collection systems, training methods and performance evidence. Each article asks what changed technically, what the evidence supports, and what it means for the cost, speed or reliability of building AI. Embodied intelligence belongs here when the story concerns the model or its training data; robot product reviews belong elsewhere.

Who we are

Frontier AI Review is an editorial publication focused on the technologies and decisions reshaping AI. We connect technical developments to operational realities, economics and public accountability.

The team

Individual editor and contributor biographies have not yet been published. For editorial enquiries, contact [email protected].

How we work

We develop original arguments using attributed evidence. Articles distinguish company claims, demonstrations and deployed capability, and explain what the available record does not establish. Read our editorial standards and corrections policy.

Photo credits

Each article credits its images in the accompanying caption. Images may come from manufacturers, institutions or other identified rights holders. A product photograph does not establish performance, and a demonstration is not evidence of routine operational use. Illustrations are identified as such. Image rights remain with their owners; publication does not imply endorsement.

For an image attribution or permissions concern, include the article address and image details in an email to [email protected].