Pricing
Two enforced quotas: concurrent missions in flight, and agents currently bound to a mission task. Idle agents do not count.
Team
For small red teams getting started with continuous AI red teaming
$799/mo
- 10 concurrent missionsin non-terminal execution at any moment
- 100 concurrent agentsbound to in-flight tasks at any moment
Org
Most popularFor security organizations running coverage across multiple AI surfaces
$4,999/mo
- 50 concurrent missionsin non-terminal execution at any moment
- 250 concurrent agentsbound to in-flight tasks at any moment
Enterprise
For platform teams operating Gibson at scale across business units
$9,999/mo
- 75 concurrent missionsin non-terminal execution at any moment
- Unlimited concurrent agentsbound to in-flight tasks at any moment
Enterprise (On-Prem & Federal)
Self-hostedRun the full platform inside your own boundary
Deploys into your own Kubernetes
Helm chart deploys the full Gibson stack into your Kubernetes cluster. No egress, no telemetry, no callbacks home.
Bring your own models
Run against your own LLM endpoints, vLLM clusters, or vendor inference deployments. Prompts and results never leave your boundary.
Bring your own secrets backend
Integrate with HashiCorp Vault, AWS KMS, or your existing PKI. Per-tenant KEK derivation and workload identity built in.
SSO and SCIM
OIDC and SAML to your IdP (Okta, Entra, PingFederate, Keycloak). SCIM 2.0 user lifecycle. Per-tenant role mapping.
Audit log streaming
Export every authentication, authorization, and orchestration event to your SIEM via syslog, Kafka, or OpenTelemetry.
Capable of GovCloud, IL5 / IL6, and FIPS 140-3 environments
Ships against FIPS-validated cryptographic modules. Compatible with AWS GovCloud (US), Azure Government, and on-prem environments accredited to IL5 / IL6. Customer holds the ATO; we provide the artifacts.
Custom retention and data residency
Pick where mission data, embeddings, and graph state live. Configurable retention windows per data class.
Dedicated forward-deployed engineers
Direct Slack or Teams channel with the engineers who build Gibson. Defined response times, scheduled office hours, custom SLAs.
Talk to the team
Quotas, pricing, and rollout shape are set per engagement.