Belgian cybersecurity company Aikido has launched an open-weight artificial-intelligence model designed for cybersecurity applications that can be deployed locally, Reuters reported on September 21.
The model, called Altar, is a compressed and customised version of Chinese AI company Z.ai’s open model GLM-5.3, according to Aikido. The company said Altar will be openly available for local deployment, allowing organisations to use AI-assisted security tools without having to send sensitive source code to external providers.
Demand for locally run security models has grown as companies become more cautious about code privacy and as cybercriminals increasingly use AI to help identify and exploit software weaknesses. Running a model inside an organisation’s own environment can reduce some data-exposure risks, although local deployment does not by itself eliminate broader security or model-governance concerns.
Reuters reported that Altar is expected to be deployed in Aikido products used by customers including Belgian bank Belfius. Aikido reached a valuation of about $1 billion in January.
The launch comes as cybersecurity companies compete to build AI tools for vulnerability detection, code review and software protection while customers weigh the benefits of cloud-based systems against greater control over sensitive internal data.