Muse Glimmer is an open model built for always-on local agents: 30B parameters, licensed under Apache 2.0, and tuned for tool use, long tasks, and failure recovery.
Why a model for always-on agents
Most models are optimized for a single answer. Agents are different: they run for hours, call tools, hit errors, and have to keep going anyway.
Muse Glimmer is trained for that loop:
- Reliable tool-calling
- Persistent state across restarts
- Self-managed memory across hours-long sessions
Where it runs
Muse Glimmer is small enough to run on a single consumer GPU or Mac, enabling use cases that range from local agents to local coding.
| Platform | How to run |
|---|---|
| Ollama | ollama run muse-glimmer |
| LM Studio | Search for Muse Glimmer in the model browser |
| Hugging Face | transformers, llama.cpp, vLLM, Colab |
| vLLM | OpenAI-compatible server |
Use it as a web agent session
You can also run Muse Glimmer as an agent session on the web. Sign in with Google and your account starts with 50 starter credits. Agentic sessions run on credits, and credit packs are one-time purchases that never expire.
