DEPLOYDatabase

Company

Reka AI

reka.ai

AI lab building multimodal and omni models, including the Rho-1 omni-reasoning model that unifies text, image, video and robot action tokens in one network.

Track Reka AI
Status
Private

Deploy Watch

Track Reka AI on Deploy.

We notify you when a verified state change lands on the record, not when the company issues a claim.

Track Reka AI

Key facts

Rho-1 (company-stated, Oct 2026)

A 19B omni-reasoning research preview, trained from scratch, that understands and generates text, images and video and takes actions.

World-language-action design (company-stated)

Robot actions and proprioception are continuous tokens; the state that predicts future camera frames also outputs the action trajectory.

Inverse Dynamics Model (company-stated)

Used with Rho-1 to infer control signals from unlabeled internet video.

Training (reported, Oct 2026)

Trained on 320 H100 GPUs over about three months.

Robotics demo (reported, Oct 2026)

A LIBERO simulation task with seven action channels and a predicted wrist-camera view. No real-robot or sim-to-real results disclosed.

Data & sources

Web sources

4

4 sources backing this record.View all →

Safety record

No incidents on record for Reka AI.

Only active incidents are counted. Retracted incidents are excluded from this summary but remain reachable at their canonical URLs.

Full safety record: incidents, sourcing, and exposure data →