Icon for Magnitude

Magnitude

Open Source Free Trial

Open-source desktop inference engine that tunes its kernels to your hardware and connects local models to your agent.

Magnitude is an open-source inference engine that ships as a desktop app for running open-weight models locally. It compiles and tunes its kernels on your own device before a model runs, on Apple Silicon, NVIDIA or AMD GPUs, or CPU only, across macOS, Linux and Windows. Concurrent agent sessions share prefix caches.

Its own benchmark on Qwen 3.6 35B A3B reports 92% faster decode than llama.cpp on an M4 Pro and 19% on a DGX Spark. It connects in one click to Codex, Claude Code, OpenCode and other agents. Free under Apache-2.0.

Pricing: Free

Hosting Self-hosted
Pricing Free, Free (open source)
License APACHE-2.0
Screenshot of Magnitude webpage

Work on Magnitude? Feature it at the top of Frameworks & Stacks.

Is your product missing?

Add it here →