Free guide · Sep 17, 2026

Run a 744B-parameter AI model with no GPU

What Colibri really needs (about 25GB of RAM), how to set it up, and the catches.

The short version

Colibri is a free, open source engine that runs huge AI models on normal computers. It keeps the core of the model in RAM and streams the rest from your SSD, so you trade GPU memory for disk space and patience.

•What you need: about 380 GB of free SSD space, 24 GB of RAM or more, and no GPU.

•What speed to expect: anywhere from one token every 10 to 20 seconds up to several tokens a second, depending mostly on your RAM and your drive. The table below shows real numbers.

•How to start: install the prebuilt release, test with a 20 GB model first, then download the big one.

What's in the full guide

  1. 01What speed to expect
  2. 02What you need
  3. 03Step 1: Install Colibri
  4. 04Step 2: Test with a small model first
  5. 05Step 3: Download GLM-5.2
  6. 06Step 4: Run it
  7. 07Step 5: Speed test your machine
  8. 08Why it works
  9. 09Troubleshooting
  10. 10Is it worth it?

From the blog

AI Tools

10 MCP servers

Stride AI Academy

AI Tools6 min read

Best MCP Servers for Claude Code: 10 We'd Set Up First

The MCP servers we'd set up first in Claude Code for client builds: what each does, when to use it, the install command and how to keep access safe.

More AI Tools guides

AI Tools

olmOCR 2: turn scanned PDFs into text your AI can read

How to pick a small scan to try, check numbers and equations, and see whether your GPU fits before you download it.

Read the guide →
AI Tools

SynthID Bio: see where the protein watermark goes and how it is detected

Where the mark goes, how they detect it, and what the marked proteins could still do.

Read the guide →
AI Tools

LFM2.5-VL DSpark: measure the real speedup on your own request

How to see where the time goes, match the two models, and compare the same request on both.

Read the guide →