AI red teaming & LLM evaluation
Gray Swan is an AI red teaming and agent assessment platform from Gray Swan. It is a commercial product offered as: SaaS, On-prem, Managed service. Pricing model: not published.
Best for
Gray Swan's red teaming draws on a crowd. It runs Arena, a public competition that pays cash prizes for new AI vulnerabilities and that the home page says involves over 15,000 adversarial researchers, and it feeds what it learns into two offers. Shade is the automated product: adversarial agents attack your model, guardrails and tools, covering jailbreaks, prompt injection, data leakage and behavior outside policy, and return reproducible cases with severity. The private AI Red-Teaming service puts hand-picked Arena top performers on a scoped engagement. Buyers can start with the product and add people, or the reverse, from one vendor that also builds evaluations for model labs. The public API reference covers its Cygnal runtime product, not Shade, and Shade setup starts with a call to Gray Swan, so plan on a sales-led rollout.
This summary is our own wording of the sourced facts on this page.
Each value links to the page that states it, with the date we checked it.
As documented by the vendor or project. The tag is our label from a controlled vocabulary.
Our editorial model, not a standard. The primary layer follows the category, except for tools whose main documented job is detection and response; other layers rest on a capability the sources document. How the model works.
| Layer | Role | How it applies | Based on |
|---|---|---|---|
| L4 Testing Testing and red teaming | Primary | Attack and evaluate models, apps and agents on purpose, before release and after every change, to find what the other layers miss. | Follows its category, AI red teaming & LLM evaluation |
The risk list is OWASP's (Top 10 for LLM Applications, 2025). The mapping is our analysis, and each row rests on a capability the sources document.
| Risk | How it applies | Based on |
|---|---|---|
| LLM02 Sensitive Information Disclosure | Shade's data leakage category tries to extract training data or sensitive information. | Shade tests data leakage, meaning attempts to extract training data or sensitive information |
| LLM01 Prompt Injection | The managed service covers prompt injection among its attack types. | The managed AI Red-Teaming service lists prompt injection among its attack types |
| LLM06 Excessive Agency | The managed service covers tool exploitation against agent deployments. | The managed service lists tool exploitation among its attack types, with scope that includes tools and retrieval systems |
More in AI red teaming & LLM evaluation, ranked within the category by Coverage Score.
NVIDIA
Probes language models and LLM apps for prompt injection, jailbreaks, data leakage and more.
Best for: Teams that want a free command-line scanner they can point at any model or LLM endpoint before release
Reviewed
Giskard AI
Generates adversarial test suites for LLM agents and RAG apps, runs them and reports failures by threat type.
Best for: Python teams that want a free library to scan an LLM agent or RAG app for prompt injection and harmful output, and to write pass or fail checks for it
Reviewed
OpenAI (Promptfoo)
Runs evals and red team scans against LLM apps from a CLI or CI pipeline and writes a vulnerability report.
Best for: Developer teams that want one open-source CLI for both prompt evals and red teaming, run locally and in CI
Reviewed
Shade is the automated product that runs adversarial agents against your deployment. The AI Red-Teaming service is a private engagement run by hand-selected Arena top performers, with an executive summary, transcripts and a remediation roadmap.
No. The pages we opened show no prices, only buttons such as Schedule a Demo and Talk to an AI Security Expert.
No public source repository is linked from the pages we opened, and the product is sold as a commercial platform and service.
Not from what is public. The Gray Swan API reference lists endpoints for its Cygnal runtime product, policies and activity export, none for Shade, and the Shade docs describe no CI integration; a Shade evaluation is set up with the Gray Swan team.
Gray Swan is closed source (proprietary).
Source Source, checked October 21, 2026: grayswan.ai (opens in a new tab) Checked
Gray Swan is offered as: SaaS, On-prem, Managed service.
Source Source, checked October 21, 2026: grayswan.ai (opens in a new tab) Checked
Pricing model for Gray Swan: Not published in the sources we checked.
Source Source, checked October 21, 2026: grayswan.ai (opens in a new tab) Checked
Status of Gray Swan: Active.
Source Source, checked October 21, 2026: grayswan.ai (opens in a new tab) Checked
We have no commercial relationship with Gray Swan.