Magma: A Foundation Model for Multimodal AI Agents

TL;DR

Magma is a versatile foundation model that integrates vision-language understanding with spatial-temporal reasoning, enabling advanced multimodal AI agents to perform tasks in digital and physical environments.

Contribution

Magma extends vision-language models by incorporating spatial-temporal reasoning and agentic capabilities through novel training with heterogeneous datasets and action grounding techniques.

Findings

Achieves state-of-the-art results in UI navigation and robotic manipulation.

Outperforms existing models on image and video multimodal tasks.

Demonstrates effective synergy of SoM and ToM for spatial-temporal intelligence.

Abstract

We present Magma, a foundation model that serves multimodal AI agentic tasks in both the digital and physical worlds. Magma is a significant extension of vision-language (VL) models in that it not only retains the VL understanding ability (verbal intelligence) of the latter, but is also equipped with the ability to plan and act in the visual-spatial world (spatial-temporal intelligence) and complete agentic tasks ranging from UI navigation to robot manipulation. To endow the agentic capabilities, Magma is pretrained on large amounts of heterogeneous datasets spanning from images, videos to robotics data, where the actionable visual objects (e.g., clickable buttons in GUI) in images are labeled by Set-of-Mark (SoM) for action grounding, and the object movements (e.g., the trace of human hands or robotic arms) in videos are labeled by Trace-of-Mark (ToM) for action planning. Extensive experiments show that SoM and ToM reach great synergy and facilitate the acquisition of spatial-temporal intelligence for our Magma model, which is fundamental to a wide range of tasks as shown in Fig.1. In particular, Magma creates new state-of-the-art results on UI navigation and robotic manipulation tasks, outperforming previous models that are specifically tailored to these tasks. On image and video-related multimodal tasks, Magma also compares favorably to popular large multimodal models that are trained on much larger datasets. We make our model and code public for reproducibility at https://microsoft.github.io/Magma.