Muse Glimmer
Meta's open agentic 30B-parameter language model that runs locally.
HN commenters like it for producing decent results while running locally at 30B params123, and for using tokens efficiently enough to finish tasks quicker than A3B.45 The most common complaint is that it runs slowly6789, and one commenter found it changes its position too easily compared with Qwen.10
Why people recommend it
Output quality · 3
Commenters report decent results locally and impressive chat.
“runs locally at 30B params. Slow for most things but it does produce decent results”
“I am getting good results with muse-glimmer running locally, with the caveat that everything runs slowly”
Show 1 more
“I am enjoying testing Muse Glimmer — it's really quite impressive on chat”
Token efficiency · 2
It uses tokens sparingly, so tasks finish faster despite slower generation.
“it uses them very efficiently, caveman style”
“I have been using Muse Glimmer for a few days instead of A3B. It gets the job done quicker than A3B despite being several times slower.”
Runs locally · 2
The 30B model runs on local hardware, including an older 32GB Mac Mini.
“runs locally at 30B params. Slow for most things but it does produce decent results”
“I am getting good results with muse-glimmer running locally, with the caveat that everything runs slowly”
Writing personality · 1
One commenter enjoys its fun, slightly snarky writing style.
“the way Muse Glimmer writes. Like a talented, slightly snarky guy who is maybe a bit of a dick but quite fun to be around.”
Caveats and dissent
Slow generation · 3
Several commenters say it is slow per token and slow for most tasks.
“I am getting good results with muse-glimmer running locally, with the caveat that everything runs slowly”
“runs locally at 30B params. Slow for most things but it does produce decent results”
Show 1 more
“I have been using Muse Glimmer for a few days instead of A3B. It gets the job done quicker than A3B despite being several times slower.”
Inconsistent stance · 1
One commenter found it flips between positions too easily and preferred Qwen for critique.
“Muse came close but what impressed me about Qwen is it’ll push back”
Often compared with
Who recommends it in their own words
Threads where it came up
- Meta Muse Glimmer – open weights 30B local coding model 2 mentions
- Astra and Fable still hack on simple variants of alignment evals from 2025 1 mention
- GLM-5.3 is now open-weight 1 mention
- Qwen 3.8 27B is excellent, but it defaults to overthinking things 1 mention
- Why does Opus 5 feel worse to work with? 1 mention
- Ask HN: Claude Code Alternatives 1 mention
Also written as muse-glimmer. Data: muse-glimmer.json. Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 7
“Muse came close but what impressed me about Qwen is it’ll push back”
“it uses them very efficiently, caveman style”
“I have been using Muse Glimmer for a few days instead of A3B. It gets the job done quicker than A3B despite being several times slower.”
“the way Muse Glimmer writes. Like a talented, slightly snarky guy who is maybe a bit of a dick but quite fun to be around.”
“runs locally at 30B params. Slow for most things but it does produce decent results”
“I am enjoying testing Muse Glimmer — it's really quite impressive on chat”
“I am getting good results with muse-glimmer running locally, with the caveat that everything runs slowly”