Colibri
No written verdict yet. Here is what commenters said.
Why people recommend it
“Colibri did that first for GLM-5.2”
“Colibri [2] which have better support for streaming weights from disk”
Show 3 more
“Check for colibri, dwarf star and flash-moe.”
“Check the Colibri and Flash-MoE projects. They’re already doing that.”
“Or from SSD using something like Colibri[1]. Not going to be quick, but at least runable.”
Caveats and dissent
“I'm aware of colibri [1], but so far I've only seen extremely slow performance.”
Often compared with
From the makers not counted in the ranking
“Also, feel free to DM me (@colibri.social) on Bluesky if you want to migrate to the Colibri PDS!”
“Right now, Colibri is meant to function for communities that are public anyway.”
“We're doing that to allow you to update your profile from within the app.”
“I'm working on a self-hostable ebook library”
“Colibri[1], a self-hostable ebook library.”
“An open source, self-hostable ebook library for yourself”
Threads where it came up
- Colibri – chat platform built on the AT Protocol for communities big and small 5 mentions
- Ask HN: What Are You Working On? (January 2026) 3 mentions
- Show HN: Open-source engine running Gemma 4 26B in 2 GB RAM on any M-series Mac 2 mentions
- Kimi-K3 Releases on HuggingFace 6/27 2 mentions
- Show HN: Running 104GB Qwen3.8-Flash-Next on 48GB Mac with at ~12 tok/s 1 mention
- GLM-5.3-Flash 1 mention
Also written as colibri. Data: colibri.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 14
“Colibri did that first for GLM-5.2”
“similar to how Colibri works”
“I wonder how it could work with something like https://github.com/JustVugg/colibri.”
“Colibri [2] which have better support for streaming weights from disk”
“Check for colibri, dwarf star and flash-moe.”
“Check the Colibri and Flash-MoE projects. They’re already doing that.”
“colibri is meant to allow running GLM 5.2 the 744B model in 32GB”
“Or from SSD using something like Colibri[1]. Not going to be quick, but at least runable.”
“I'm aware of colibri [1], but so far I've only seen extremely slow performance.”
“Someone did this exact thing recently”
“Already is possible.”
“you should check out the project this was based on: https://github.com/JustVugg/colibri”
The first assumption has been long disproved since multiple full scale Discord data leaks. If it's a public server, it can be scraped. https://www.malwarebytes.com/blog/news/2024/04/billions-of-scraped-discord-messages-up-for-sale
“This implementation is almost entirely vibe-coded”