Muse Glimmer

Meta's open-weight multimodal model, made to run on your own hardware.

Official site Pricing page

Open weights; you pay for your own hardware. Not on Model API. Checked 2026-10-14.

Quick start

Download the weights, serve them with vLLM or llama.cpp, and point your app at the local OpenAI-compatible endpoint.

Good and not so good

• Runs locally via vLLM, SGLang or llama.cpp
• No per-token fees
• You run and scale the hardware

Use cases with Muse Glimmer

No use cases yet. Add an agent use case.

Alternatives