Meta Open-Sources 30B Parameter Agent Model MuseGlimmer, Runs Locally on 24GB VRAM
AI寒武纪 · wechat · 2026-08-10
Meta announced the open-source release of MuseGlimmer, a 30B parameter Agent model under the Apache 2.0 license, with weights now available on HuggingFace. Designed specifically for agentic tasks, the model leverages 4-bit quantization to compress its size below 20GB, allowing it to run smoothly on local hardware with just 24GB of VRAM without compromising Agent capabilities.
The model features core abilities including end-to-end task completion, tool calling, multi-step reasoning, automatic failure recovery, and multimodal understanding. It demonstrates strong performance on benchmarks like DeepSearchQA and SWE-Bench. Furthermore, it integrates speculative decoding technology to significantly accelerate generation speed for long reasoning chains. The model is currently optimized for popular deployment tools like Ollama and vLLM.
Related event: Meta Open-Sources 30B Muse Glimmer for On-Device Agents(108 posts)→
More from coding & agent
- Opus 5.5 + Three.js demo stuns with water physics, fish and blossoms — nikola_mr64990 · 2026-10-16
- Karpathy's knowledge engine workflow: Claude Code turns raw notes into a self-updating wiki — nikola_mr64990 · 2026-10-16
- Sierra and Meta unveil Personal Agent Protocol with Zapier as launch partner — sujingshen · 2026-10-16
- ArcKit adds 102 AI-narrated short films for commands across 14 community overlays — mcraddock · 2026-10-16
- Personal agents live or die on chaining tasks, not single-task ability — sujingshen · 2026-10-16
- Dev laments durable coding agent crumbling on restarts with long subagent sessions — lucasmeijer · 2026-10-16