Our Models

Reka has developed three models to fit a variety of use cases. All of our models are fully multimodal, allowing for image, video, audio, pdf and text input.

We believe you should determine how our models are deployed and your data security and privacy are paramount. Therefore, our models can be deployed in the cloud, on-premises, or on-device.

Our models can be customized, and we allow for fine-tuning of our models by you or with our help.

Edge

  • 7B parameters

  • Fully multimodal

  • Succinct and powerful

  • Multilingual in 32 languages

  • 128K context length

  • Appropriate for on-device deployment (smaller variants available)

Flash

  • 21B parameters

  • Fully multimodal

  • Robust and muti-faceted

  • Multilingual in 32 languages

  • 128K context length

  • Efficient for on-premises deployment

Core

  • Larger than Flash

  • Fully multimodal

  • Detailed and comprehensive

  • Multilingual in 32 languages

  • 128K context length

  • Ideal for customization with specific data or for specific capabilities