Changelog

What's new in Cyrion AI. Entries list customer-facing changes only, newest first.

September 2026

Cyrion AI launches with simple, credit-based plans.

Lite at $99 per month with 200 credits, Hacker at $999 per month with 3,000 credits, Red Team at $1,999 per month with 8,000 credits, and Enterprise with custom terms.
Credits are metered by actual usage, and scans pause instead of failing when a balance runs out.
One-time credit packs for extra headroom on paid plans.
The Free plan is now a full demo: a complete demo scan and a sample report.

Added GPT-6 Astra and GPT-5.6 Sol, Terra and Luna.
Added Claude Fable 5.1, Claude Opus 5 and Claude Sonnet 5.
Added Gemini 3.8 Flash and Gemini 3.1 Pro, DeepSeek V4.1 Flash, Grok 4.6, Qwen3.8, Kimi K3, GLM-5.3 and MiniMax-M3.
Most models offer 1M-token context windows.
Choose the model per scan, or assign different models to the Large, Medium and Small work tiers.
Lite gets a curated menu of fast, cost-efficient models.

Four scan modes: Quick, Standard, Aggressive and Thorough. Lite includes Quick and Standard, Hacker adds Aggressive and Thorough.
Ten target types including web applications, repositories, IP ranges, Android and iOS apps, AWS, Azure, GCP and AI applications.
Registered targets: scans are bound to a target you register, with limits by plan.
Engagement agreement signed before every scan, relaunch and clone.
Pre-scan health check catches unreachable targets and invalid credentials before credits are spent.
Authenticated testing with basic, bearer and API key headers, form login flows, custom cookies and TOTP.
Compliance flags to keep sensitive data out of findings.
Scheduling from the wizard, with a Calendar view of upcoming scans.
Stealth mode routes traffic through Tor on Hacker and above.

Already part of Cyrion AI at launch:

Validated findings with proof of concept, CVSS score, reproduction steps and remediation guidance.
Browser-captured screenshots and visual proof on Hacker and above.
One-click Retest for any finding.
Attack Chain view that shows how findings connect.
Triage, comments, favorites and a Taskboard for remediation tasks.

Penetration Test, OWASP, Bug Bounty, Executive, SOC 2, ISO 27001, PCI DSS, Compliance and AI Red Team report types.
PDF, DOCX and JSON export, scheduled reports, a report library and a content editor.
29 compliance frameworks across Lite, Hacker and Red Team.
Custom templates and branding on Red Team, and white-label reports on Enterprise.

GitHub, Jira, Slack, Microsoft Teams, Discord and Burp Suite Enterprise, all connected from inside the app.
MCP servers for threat intelligence, including NVD, EPSS, CISA KEV, MITRE ATT&CK and Shodan.
Secrets Vault and API keys for secure, scripted access.
A REST API covering scans, findings, reports, integrations and team management. See the API Reference.

Cloud scanning for AWS, Azure and GCP.
AI Red Team for LLM and AI application testing, with Adaptive attack mode, hunt modes, browser targets and Target Intel.
Team management with 3 seats on one shared credit balance, groups and custom roles.
Uncensored AI models and extended compliance reports.