Skip to content
ThreadOut
Alexandr Wang@alexandr_wang · Aug 10, 2026

1/ big announcement today: we will be releasing an open weight version of muse spark 1.2 soon. we also are releasing…

6 tweets1 min read♥ 9.5Koriginal

Summary

A company is releasing Muse Glimmer, a 30 billion parameter open-weight agentic AI model that runs efficiently on 24GB of VRAM and can operate as a fully capable agent with planning, tool calls, and failure recovery. The model is available on Hugging Face and compatible with multiple inference frameworks.

Summarized by ThreadOut AI from the full thread. May miss nuance — read the thread below.

  1. #1

    1/ big announcement today: we will be releasing an open weight version of muse spark 1.2 soon. we also are releasing muse glimmer, a 30B agentic model with open weights under apache 2.0. muse glimmer can run on 24GB of VRAM without losing agentic reliability. 🧵

  2. #2

    2/ just like much larger models, muse glimmer can operate as a fully capable agent via planning, tool calls, checking its own results, and failure recovery.

  3. #3

    3/ muse glimmer was developed with its own architecture and recipe, optimized for its size and agentic performance requirements.

  4. #4

    4/ to achieve maximum memory efficiency, we quantize model weights to ~4-bit, getting the language model under 20GB with room for the kv cache, perception encoder, and drafter alongside it. a dflash drafter proposes blocks of tokens the main model verifies in parallel, so it stays responsive.

  5. #5

    5/ weights on hugging face now. running this week through ollama, LM Studio, vllm, sglang, together, fireworks, and openrouter, with llama.cpp, MLX, and executorch. evaluated under meta's advanced AI scaling framework and cleared for open-weight release.

  6. #6

    6/ that's the short version. the full write-up, including evals and the training details, is here: research.meta.ai/blog/introduci…