Fine-Tune Muse Glimmer 30B with LoRA or Full-Parameter on Fireworks

Melvin Vivas · X post · 2026-08-19 · Open on X

Topics: Fine-tuning & Model Customization, AI Agents, Tool Use & MCP · Level: advanced

Summary

Muse Glimmer 30B is now on Fireworks' Dedicated Training API for both LoRA and full-parameter fine-tuning. It is a U.S.-developed open-weight model said to be strong at agentic work, tool use and long-horizon reasoning.

Key points

• Fireworks' Dedicated Training API supports LoRA and full-parameter fine-tuning.
• Muse Glimmer 30B is an open-weight, U.S.-developed model.
• It is described as one of the strongest models of its size for agentic work and reliable tool use.
• LoRA is the cheaper option; full-parameter tuning updates all the weights.

Resources mentioned

• Fireworks AI Dedicated Training API · tool · docs.fireworks.ai · paid
  Fireworks API for LoRA and full-parameter fine-tuning of open models.
• Muse Glimmer-30B · tool · developer.meta.com · free
  Open-weight 30B model aimed at agentic work, tool use and long-horizon reasoning.
  Also in: Meta Muse Glimmer-30B: Open Weights Model That Beats Qwen3.6 37B on Agentic Tasks (Melvin Vivas on X · notes)