Best AI Evals

Vendor profileChecked Nov 8, 2026

Arthur

Arthur (arthur.ai) · arthur.ai · Last checked:

  • Hosted SaaS platform
  • Open-source evals engine (Arthur Engine, MIT)
  • Self-hosted or VPC deployment (Enterprise)

Company status

Active, privately held.

Sources 1, 2, 3 · checked 2026-11-08

Capabilities

  • Agent discovery, governance, audit logs, and lineage
  • Guardrails and runtime detection with anomaly alerting
  • Continuous evals; open-source Arthur Evals Engine with PII, sensitive data, custom LLM, and regex rules
  • Prompt experimentation through continuous evals across the agent lifecycle

Sources 1, 2, 3 · checked 2026-11-08

Who it tends to fit

  • Security, audit, and governance teams that want agent inventory alongside evals
  • Teams that want an MIT-licensed evals engine they can deploy themselves

Sources 1, 2, 3 · checked 2026-11-08

Pricing

Free: $0/mo (monitoring for up to 4 use cases, unlimited seats). Premium: $60/mo (monitoring for up to 100 use cases). Enterprise: custom (dedicated and managed VPC options).

Sources 1, 2, 3 · checked 2026-11-08

Published certifications

None found / not verified. This is not a finding that Arthur lacks certifications; ask the vendor for documentation.

The pricing comparison table has a SOC 2 row, but no certification statement was found on a fetched page; /security returned 404 on 2026-11-08. Ask the vendor for its report.

Listed as stated by the vendor. Scope can differ by product or subsidiary; request current reports.

Sources 1, 2, 3 · checked 2026-11-08

Limitations to check

  • Product positioning centers on agent governance and security; confirm the depth of offline eval and experiment features for your use case
  • No certification statement found on official pages

Sources 1, 2, 3 · checked 2026-11-08

Sources

  1. Free $0/mo, Premium $60/mo, Enterprise custom with dedicated and managed VPC options; open-source Arthur Evals Enginewww.arthur.ai/pricingChecked 2026-11-08
  2. Homepage: agent security and governance, guardrails, continuous evalswww.arthur.ai/Checked 2026-11-08
  3. arthur-engine repository is MIT licensed, topics include evaluation, guardrails, tracinggithub.com/arthur-ai/arthur-engineChecked 2026-11-08

Report a correction: [email protected]. Facts are re-checked at least every 90 days.