Aegis A1 · Security Intelligence Model · Training

The first serious AI model for offensive security.

32 billion parameters. 2.1M+ training records. 30 security domains. Built for operator workflows — not general-purpose chat.

Join the Waitlist

Parameters

32 Billion

Training Records

2.1M+

Security Domains

Training Hardware

AMD MI300X — 192GB HBM3

Status

Training in progress

About the Model

Built from the ground up for security operators.

Aegis A1 is not a general model with security prompts bolted on. It is a domain-specific model trained on curated offensive-security data — structured multi-turn conversations across 30 topics, designed to develop deep technical reasoning.

The base model was selected for its lack of refusals. The fine-tune adds operator-grade reasoning depth, pentest-specific workflows, and the Aegis identity. The result is a model that reasons like a senior operator, not a compliance filter.

Training Coverage

30 offensive-security domains.

SQL Injection

Cross-Site Scripting

SSRF Exploitation

Authentication Attacks

File Upload Exploitation

XXE & Deserialization

IDOR & Access Control

API Security

GraphQL Attacks

OAuth & JWT Attacks

Port Scanning

SMB Exploitation

Active Directory

VPN & Firewall Evasion

DNS & Subdomain Attacks

Wireless Attacks

AWS Misconfigurations

Container & Kubernetes

Buffer Overflow

CVE Exploitation

Linux Privilege Escalation

Windows Privilege Escalation

Post-Exploitation

Lateral Movement

Malware Analysis

Reverse Engineering

Social Engineering

Report Writing

Network Recon

OSINT