Reka AI builds frontier multimodal AI models capable of understanding and reasoning across text, images, video, and audio. Its model family, including the Flash and Core models, is offered via API and on-premise deployment, targeting enterprises that need reliable multimodal reasoning with efficient inference economics rather than only frontier-scale text models.
Reka's technical bet is efficiency at the frontier: delivering competitive multimodal performance at lower inference cost and with deployment flexibility, which matters for enterprises that cannot send sensitive data to third-party clouds or that need predictable unit economics at scale.
The company was founded in 2022 by researchers including Yi Tay, Dani Yogatama, Cyprien de Masson d'Autume, Mikel Artetxe, and Qi Liu, bringing together experience from Google DeepMind, Google Brain, Meta FAIR, and Baidu. Chief Scientist Yi Tay and co-founder Dani Yogatama lead the research organization.
In July 2025 Reka raised $110 million in a round backed by NVIDIA and Snowflake, reaching a valuation above $1 billion and unicorn status. Reka's total funding across its rounds is reported at roughly $168 million, with the latest round structured as a Series B.
Reka differentiates through native multimodality, efficient inference, and deployment options that include on-premise and private environments. It competes with larger frontier labs and enterprise model vendors, positioning itself for organizations that want strong video and image understanding without committing solely to the largest closed text models.