Gateco logo
ContactGet Started
Blog/performance

performance

2 posts

engineeringperformancetransparency

Our Site Said "<25ms Overhead." We Could Not Prove It. Measuring Found Three Bugs.

We went looking for the benchmark behind our own published latency figure and could not find one. Measuring properly started at 302ms p95, surfaced three product bugs, and ended at a public, reproducible 16ms p50 / 21ms p95.

August 23, 2026·8 min read
engineeringperformancearchitecture

Measuring Gateco Policy Overhead: 16ms p50, 21ms p95

How much latency does an authorization layer add to RAG? The measured answer, with a public benchmark: 16ms p50 and 21ms p95 policy overhead, and what drives variance across connectors.

April 18, 2026·6 min read
← All posts
Gateco

Permission-Aware Retrieval for AI Systems

21ms p95 measured overhead•Fail-closed by default
Get Started, Free
Available onAzure MarketplaceAvailable in theOkta Integration Network

Product

  • Features
  • Pricing
  • Integrations
  • Architecture
  • MCP Server

Solutions

  • RAG Authorization
  • Permission-Aware Retrieval
  • RAG Access Control
  • For Security Teams
  • Customers
  • EU AI Act

Compare

  • All comparisons
  • vs Cerbos
  • vs pgvector RLS
  • vs Microsoft Purview
  • vs Glean
  • vs Pinecone RBAC
  • vs Oso
  • vs OPA / Rego
  • Build vs buy

Resources

  • Documentation
  • SDKs
  • Guides
  • Blog
  • Videos
  • Changelog
  • Roadmap
  • Design Partners
  • About

Compliance

  • Security
  • Trust Center
  • HIPAA-Compliant RAG
  • GDPR-Compliant RAG
  • SOC 2
  • GDPR
  • HIPAA
  • EU AI Act
  • Privacy Policy
  • Terms of Service

© 2026 Gateco. All rights reserved.

security@gateco.aienterprise@gateco.ai