Suite

CSI Cybersecurity Superintelligence

CSI is the complete Cybersecurity AI suite. Not a scaffold, not a model, not an agent — all of them. Six tightly integrated layers (LLMs, scaffolds, datasets, agents, steering and benchmarks) shipped as one product, designed to run on-prem.

From the first byte of training data to the last shell command an agent executes, CSI is what runs the whole pipeline.

Buy CSI
Co-funded by the European Innovation Council (EIC)
SUITECSI MODELalias2-mini RUNSon-prem
DEMO
// The suite

Six layers. One product.

CSI is not just our scaffold combiner. CSI is the entire Cybersecurity AI stack: every layer designed to reinforce the next, all shipped under one licence and one routing proxy.

01

LLMs

Cybersecurity-specialized models — alias3, alias2, alias2-mini. World's best on Cybench.
02

Scaffolds

CSI scaffold combiner over Claude Code, Codex, Mistral, CAI & GCAI — with a unified routing proxy.
03

Datasets

18.07 TB of expert security trajectories from 123 countries. EIC-co-funded sovereign corpus.
04

Agents

15+ specialized cybersecurity agents — defender, red team, APT, forensics, robot defender, custom.
05

Steering

Activation steering & abliteration — refusals from 59% to 1% on offensive tasks.
06

Benchmarking

11× faster, 156× cheaper than humans — the dashboard of progress. Every layer of CSI measured continuously.
// Architecture

How it all fits together

Every layer of CSI builds on the one below it — from open research at the foundation, up to the agents that operate in the field.

1 Research 25+ open research publications
↓
2 Scaffolds CSI scaffold combiner · CAI framework
↓
3 Datasets 18.07 TB · 25.7M prompts · 224,766 sessions
↓
4 LLMs alias3, alias2, alias2-mini, alias1, alias0
Modifier Steering activation steering, abliteration
↓
5 Agents Defender, Red Team, APT, Forensics, …
Measurement Benchmarks Cybench · CAIBench · A&D CTFs

Open-source scaffolds — CSI integrates the best open-source scaffold, CAI, the scaffold created by Alias Robotics, supporting 300+ LLM models with built-in security tools, agent-based architecture and guardrails protection. Free for research purposes.

Github → Join the community →

CSI PRO

Monthly

€ 350 /month
With professional support
  • LLMs — Unlimited* tokens with alias2-mini
  • Scaffolds — CSI meta-scaffold (Claude Code, Codex, Mistral, CAI, GCAI)
  • Commercial license — use at desire
  • GDPR & NIS2 compliant
  • Transactional updates via OS-level virtualization
Buy now

CSI PRO

Yearly

BEST VALUE
€ 3,990 /year
Everything in Monthly · 5% discount
  • LLMs — Unlimited* tokens with alias2-mini
  • Scaffolds — CSI meta-scaffold (Claude Code, Codex, Mistral, CAI, GCAI)
  • Commercial license — use at desire
  • GDPR & NIS2 compliant
  • Transactional updates via OS-level virtualization
  • Agents — quarterly consulting on agent design & deployment
  • Benchmarking — quarterly consulting on measurement & reporting
Buy now

CSI On-Premise

Sovereign · Gov & Critical Infra

Custom
All six layers, on your hardware, on demand
  • LLMs — full alias family on-prem, air-gapped (incl. flagship alias2 / alias3)
  • Scaffolds — CSI meta-scaffold + CAI framework, self-hosted
  • Datasets — sovereign training corpus access & custom fine-tuning
  • Agents — bespoke agent design & deployment
  • Steering — abliteration & activation steering for your domain
  • Benchmarking — private benchmark suites, audit logging & forensics
  • GDPR & NIS2 compliant
  • Transactional updates via OS-level virtualization
Contact Sales

CLI agents powering CSI

CAI
Claude Code Claude Code
Codex Codex
Mistral Vibe Mistral Vibe

CSI is powered by alias models — security-specialized LLMs hosted by Alias Robotics in EU-compliant infrastructure. Need full data sovereignty? Deploy alias models on-premise for air-gapped, private operations with the same capabilities.

Explore models →

All prices exclude VAT. Annual contracts available with volume discounts.

assuming good usage, as per the license terms.

How security teams operate with CSI

Discover real
exposure

Identify true attack paths, external exposure and hidden adversarial opportunities across systems, environments and interconnected digital assets.

Validate security assumptions

Deploy agents that challenge systems, reproduce attacker behavior, and confirm whether protections hold under realistic adversarial conditions.

Secure development workflows

Embed security reasoning into engineering pipelines to continuously analyze code, logic, and runtime behavior before vulnerabilities propagate.

Maintain security evidence

Continuously collect, validate and organize security evidence aligned with regulatory requirements, internal controls and operational assurance needs.

Stress human & product surfaces

Simulate attacks against people, applications, APIs, devices and cyber-physical systems to uncover risk beyond traditional infrastructure boundaries.

Performance validated in adversarial environments

Three results that summarise where CSI stands today — against humans, against frontier models, against the world's best CTF teams.

// 01 · vs Human hackers
11×
FASTER
than the best human hackers
156×
CHEAPER
than the best human hackers

Source: CAI paper

// 02 · Multi-scaffold > single scaffold
The best harness is the combination

Holding the model fixed at alias2-mini, no single scaffold dominates Cybench. Combining heterogeneous scaffolds under CSI's Blackboard protocol beats every individual scaffold.

CSI::Claude
15/33
CSI::Codex
15/33
CSI::Mistral
10/33
CSI::GCAI
10/33
CSI::CAI
7/33
Union — ∪ all scaffolds
17/33
Parallel race — no-comm
17/33
Blackboard — cross-write
19/33

Source: Mayoral-Vilches et al. (2026). Towards Cybersecurity SuperIntelligence (CSI): What's the best harness for cybersecurity? arXiv:2605.28334 · Cybench 33 challenges, pass@1.

// 03 · Live international CTFs
Top of the leaderboard, worldwide

2025 saw Cybersecurity AI compete head-to-head against the best human teams on real, public CTFs.

Neurogrid CTF Rank #1 $50,000 prize · 41 of 45 flags · 155 teams
Dragos OT CTF Rank #1 peak 37% faster velocity · >1,200 teams · OT
HTB AI vs Human Rank #1 AI Top 20 Global · 19/20 flags · 163 teams
UWSP Pointer Overflow 5.2 /hour Late entry (54 days) · #21 final · 635 teams

Source: World's Top AI Agent for Security CTF

See full benchmarks →

Start operating cybersecurity workflows with CSI

Try CSI PRO → See real deployments →