Best AI Red Teaming CompaniesAn independent guide to AI red teaming vendors

Compare vendors / Gray Swan AI

Gray Swan AI

AI security lab: agent red teaming, lab evals and the Arena

Human-led engagementsAutomated platformAgents and toolsFrontier model safetyRuntime guardrailsCrowdsourced

Gray Swan was founded out of Carnegie Mellon by Matt Fredrikson and Zico Kolter. It sells Shade, an automated red teaming agent for pre-deployment and CI testing, Cygnal for runtime enforcement, scoped red teaming assessments, and private red teaming for AI labs. Its Arena hosted the 2025 UK AISI agent challenge, about 1.8 million attacks on 22 frontier agents, and relaunched in Oct 2026 with a $300K indirect prompt injection challenge.

Best for: Testing agents and models against what a large crowd of red teamers finds.

$40M Series A (May 2026)

  • – Maps to OWASP
  • – Maps to MITRE ATLAS
  • – Maps to the EU AI Act
  • ✓ SOC 2 or ISO 27001
Visit Gray Swan AI

Facts checked November 9, 2026 against the sources listed below. Gray Swan AI does not pay to appear here.

1 Legal nameGray Swan AI
2 HeadquartersPittsburgh, PA, USA
3 FoundedNot disclosed
4 TeamGrew from about 15 to 66 staff in 2026 (Fortune)
5 What they testLLM apps, Agents and tool use, Prompt injection and jailbreaks, Foundation models, Bias and harms, Frontier safety (CBRN, cyber)
6 FrameworksNot disclosed
7 CertificationsSOC 2 Type II, Cyber Essentials
8 Published pricesNot disclosed
9 Customers and partnersThe UK AI Security Institute co-ran its 2025 Agent Red-Teaming Challenge and co-wrote the results paper. Anthropic, OpenAI and Google DeepMind sponsored Arena challenges. Says Snowflake is a partner and investor. Homepage logos do not say which are customers.

Where we recommend them

Worth knowing before you sign

  • Homepage logo wall mixes sponsors, partners and customers without saying which is which.
  • Product pages name no compliance frameworks.
  • Says its benchmarks are cited in 11 frontier model system cards; that is a vendor claim.

Go to Gray Swan AI's site

Sources

  1. Vendor: homepage
  2. Vendor: trust center
  3. FinTech Global: $40M Series A
  4. Fortune: Arena 2.0
  5. UK AISI: agent competition results
  6. arXiv: Security Challenges in AI Agent Deployment

Work at Gray Swan AI? Send corrections with a source to editors@bestairedteamingcompanies.com. We fix verified errors within five business days.

Compare with

  • Mindgard AI red teaming firm: human-led tests plus an automated platform
  • HiddenLayer AI security platform with automated red teaming and model scanning
  • Trail of Bits Security research consultancy with an AI/ML assurance practice
  • Pillar Security Agent security platform with agentic red teaming
  • Straiker Agent security platform with autonomous red teaming (Ascend AI)