Reka
Multimodal models that reason across text, image, and video
Reka is profiled here as a LLM tool for engineering teams. Read about features, pricing, and how it compares to related options in the tools directory.
Description
Reka is a multimodal AI company founded in 2022 by researchers from DeepMind, Google, and Meta, with Dani Yogatama as chief executive. Its Core, Flash, and Edge models are built to understand and reason across text, images, audio, and video, with a focus on efficient training and serving at lower compute cost. Reka offers the models through an API and integrations such as Snowflake Cortex, and it has published open weights for its Flash models for teams that want local deployment.
Key Capabilities:
Native multimodal models that process text, images, audio, and video together
A model range spanning Core, Flash, and Edge for different cost and latency needs
Open-weight Flash models for self-hosted and on-device deployment
An API and platform for building multimodal applications
Long context windows and multilingual support across many languages
Agentic products for visual search and document research
Alternative tools
- GraphRAG
Graph-based retrieval-augmented generation from Microsoft Research
- LLM Guard
Open-source security toolkit for LLM interactions
- Llama Guard
Open safeguard model for classifying LLM inputs and outputs
- Martian
Model router that optimizes cost and quality per request
- Cloudflare AI Gateway
A gateway for caching, routing, and observing AI requests
- Databricks
Lakehouse platform unifying data engineering and AI