Summary

Aikido Altar is an open-weight security model built for defensive intelligence inside controlled infrastructure. It compresses a GLM-5.3-derived model with quantization and expert pruning, aiming to retain useful reasoning while reducing deployment size.

Key Features

Defensive security reasoning.
Open-weight distribution.
GLM-5.3-derived model.
Quantized weights.
Expert pruning.
A reported 328 GB model footprint.
On-premises deployment orientation.
Air-gapped security use cases.

About Aikido Altar


Security teams can investigate locally hosted intelligence for sensitive code and on-premises pentesting rather than sending that context to a third-party inference service. Core capabilities include Defensive security reasoning, Open-weight distribution, GLM-5.3-derived model, and Quantized weights.


Public code or model weights are linked from the project, with setup instructions for running or adapting the system. Additional characteristics include Expert pruning, A reported 328 GB model footprint, On-premises deployment orientation, and Air-gapped security use cases.

Get more likes & reach the top of search results by adding this button on your site!

Embed button preview - Light theme
Embed button preview - Dark theme
TurboType Banner

Subscribe to the AI Search Newsletter

Get top updates in AI to your inbox every weekend. It's free!