Reka AI

Reka describes Shutterstock using its models for automated metadata tagging of its multimedia library, with a named CEO quotation supporting the content-enrichment collaboration.

Reka describes Orange Village Police Department in Ohio deploying Reka Vision over its existing video-management and camera infrastructure for natural-language footage search. It also identifies Turing Video's Guardian AI as powered by Reka Vision. This is a named deployment distinct from Shutterstock metadata enrichment; outcome percentages on the page are supplier-reported and include inconsistent case-resolution labels, so no precise percentage is relied upon.

Reka provides cloud models, video-processing infrastructure and data infrastructure for physical-world AI, but public evidence of broad production deployment remains earlier than leading global platforms.

Reka Labs develops world-language-action and omni reasoning models; Rho-1 is a 19B model that sees, reasons, renders and acts in one shared state, while Reka Edge brings open-weight vision-language intelligence to real-time edge and physical-AI systems.

Reka develops omni models for reasoning across multiple data types in real time and is focused on foundational intelligence for robotics, wearables and physical AI.

Reka develops omni models for reasoning across multiple data types in real time and is focused on foundational intelligence for robotics, wearables and physical AI.

Reka announced on July 22, 2025 that it had secured $110 million in investment backed by new and existing investors including NVIDIA and Snowflake.

Reka reports 50+ researchers and engineers from Google DeepMind and Meta and expanded in 2026 through its combination with Moonvalley.

Reka reports 50+ researchers and engineers from Google DeepMind and Meta and expanded in 2026 through its combination with Moonvalley.

Reka Cloud supports multimodal model customization with LoRA and reinforcement learning, serverless and dedicated GPU inference, OpenAI-compatible APIs, H100/H200/Blackwell infrastructure and customer-managed GPU clusters, demonstrating skills spanning training, optimization, ML systems and deployment.

Multimodal foundation-model, video and physical-AI skills transfer broadly across robotics, wearables, media and frontier-model companies.

Multimodal foundation-model, video and physical-AI skills transfer broadly across robotics, wearables, media and frontier-model companies.