Reka AI has introduced Rho-1, an advanced omni-model featuring 19 billion parameters capable of processing text, images, video, and robot control within a single neural network. This integration across multiple modalities marks a significant step in AI development.
According to The Decoder, Rho-1 was trained using 320 H100 GPUs over approximately three months, highlighting the considerable computational resources invested in its creation. This level of scale and efficiency positions Rho-1 as a notable innovation in the AI landscape.
For Japanese markets, where AI applications in robotics and multimedia are rapidly expanding, Rho-1’s multi-modal capabilities could accelerate advancements in automation and data processing across sectors including FX and equities trading.