Multimodal AI startup Reka AI secured $110 million in new funding. Nvidia and Snowflake backed this latest investment. Consequently, this round triples Reka AI's valuation to $1 billion.
Reka AI, founded in 2022, develops multimodal AI models. These models focus on ultra-efficient training and low-cost inference delivery.
Specifically, the company offers Reka Flash, a flagship multimodal model understanding video, images, text, and audio. Reka Flash is a compact 21 billion parameter large language model.
Reka AI's Multimodal Platform Expansion
Reka Vision and Reka Research are Reka AI's agentic AI platforms. These platforms provide complex question answering and visual search capabilities. Moreover, Reka AI recently released Reka Flash 3.1, demonstrating strong performance against similarly sized models from competitors like OpenAI. The company competes with other leading AI developers, including Meta and Anthropic.
The company also developed Reka Quant, a method for almost lossless quantization. This reduces model memory footprint significantly. Therefore, Reka Quant allows running models with fewer resources. Reka AI intends to use the investment to accelerate technical development. Furthermore, it will scale its multimodal platforms for wider enterprise AI adoption.
Reka AI has grown from 20 to 50 employees over the past year. Additionally, the company develops tools assisting businesses with AI interfaces and application logic. This funding round follows a previous $50 million raise, which valued the company at approximately $300 million.
