Skip to content

GLEE

A general object foundation model for images and video

Visit official site

What glee is

The model is described as a foundation model that unifies object localization and recognition across media.

What you can do with it

  • Locate objects in an image
  • Follow objects through a video
  • Apply the model to perception tasks

What to watch out for

  • Multimodal inputs are supported
  • Zero-shot transfer is a stated ability
  • Different supervision levels are combined

Pros & cons

✓ What we like

  • Zero-shot transfer and generalization are listed
  • Scalability and robustness are stated strengths

! What to watch out for

  • Scope is object perception across media
  • Joint training draws on varied supervision

FAQ

What tasks are covered?

Detection, segmentation, tracking and recognition are listed.

How is it trained?

The material says data sources with different supervision levels are jointly trained.

Last reviewed: 2026-11-15

More LLM API platform tools

View all →

How we review