Skip to content

LTM

A model with a 100M token context window

Visit official site

What ltm is

The model is described as a very long context system from Magic that brings large code, docs and libraries into inference.

What you can do with it

  • Feed whole codebases as context
  • Build logic circuits during inference
  • Explore architecture with hash chains

Getting started

  • Visit the Magic site or GitHub page
  • Read the technical documentation
  • Configure the model for a codebase

Pros & cons

✓ What we like

  • Works with Google Cloud and NVIDIA GB200 NVL72
  • Builds complex logic circuits at inference

! What to watch out for

  • Technical documents need to be read first
  • The model must be configured for a given codebase

FAQ

How long is the context?

The material states up to 100M tokens.

Who is behind it?

The material names the Magic team.

Last reviewed: 2026-11-15

More AI coding tools tools

View all →

How we review