microsoft/Magma
A foundation model enabling AI agents to perceive, reason, and act across vision, language, and other modalities.

Velocity · 7d
+3.4
★ / day
Trend
→steady
star history
Magma is a foundation model designed for multimodal AI agents, published at CVPR 2025. It supports a broad range of agent capabilities including visual perception, language understanding, and action execution across diverse tasks. The model is released on Hugging Face and Azure AI Foundry, and provides tooling for agent deployment including a Gradio-based UI for interactive agent control.