Galileo

What is Galileo?

Galileo is an observability and evaluation platform for teams shipping LLM applications and AI agents. You build datasets from synthetic, development and live production data, add subject matter expert annotations as ground truth, and run more than 20 out-of-the-box evals for RAG, agents, safety and security alongside custom evaluators. Its auto-tune feature adjusts metrics from live feedback, an insights engine analyzes agent behavior to point to failure modes, and offline evals can be promoted into runtime guardrails. For scale, the Luna-2 small models, fine-tuned from Llama on 3B and 8B sizes, are designed to run evaluators with lower latency and cost than LLM judges, and are available on the Enterprise tier. Galileo can run as SaaS, in a virtual private cloud or on-premises. The Free plan costs $0 and includes 5,000 traces per month, Pro is $100 per month for 50,000 traces, and Enterprise is priced by contacting sales.

Use cases

• Evaluate LLM app quality before release
• Monitor production AI traffic for failures
• Turn evals into runtime guardrails
• Build ground-truth datasets from live data
• Collect expert annotations on model outputs
• Find failure modes in agent runs
• Fine-tune custom evaluation metrics

Who it's for

• AI engineers
• ML platform teams
• Enterprise AI teams

Why use Galileo

• Evals tuned from your own feedback fit your domain better than generic benchmarks.
• Moving the same evals from pre-release tests to production guardrails avoids maintaining two separate systems.
• Compact Luna-2 evaluators make scoring all production traffic more affordable than an LLM judge, according to the vendor.
• A free plan with unlimited users and unlimited custom evals lets small teams start without a sales call.
• Hosted, VPC and on-premises deployment options suit regulated enterprises.

Galileo features

• Datasets from synthetic, development and production data
• Subject matter expert annotation capture
• 20+ out-of-the-box evals for RAG, agents, safety and security
• Custom evals with auto-tune from live feedback
• Insights engine that identifies failure modes and suggests fixes
• Luna-2 small models for low-latency evaluation on Enterprise
• Real-time guardrails on Enterprise
• Role-based access control with admin, user and read-only roles
• Deployment as SaaS, virtual private cloud or on-premises

Galileo integrations

• MCP
• GPT-4o

How to get started with Galileo

1. Create a free account or book a demo, then connect your LLM app or agent so traces flow into Galileo.
2. Build datasets from synthetic, development and live data and add expert annotations as ground truth.
3. Apply out-of-the-box evals or create custom ones, then give feedback so auto-tune improves the metrics.
4. Review failure modes in the insights view, and on Enterprise promote proven evals into guardrails that score production traffic.

Galileo limitations

• Free plan is capped at 5,000 traces per month
• Real-time guardrails, Luna-2 and groups are limited to paid or Enterprise tiers
• Full RBAC, including the Manager role, requires a paid plan
• Pro pricing scales with the number of traces
• Luna-2 needs L4 or higher GPUs and about 4,000 samples to fine-tune for a custom use case

Galileo pricing

Yes, Galileo has a free plan. Paid plans start at $100 per month, according to the official pricing page on 24 Oct 2026.

Galileo plans

• Free

  $0/month

  • 5,000 traces per month
  • Unlimited users
  • Unlimited custom evals

• Pro

  $100/month

  billed yearly option; pricing scales with traces

  • 50,000 traces per month
  • Standard RBAC
  • Advanced analytics and insights
  • Dedicated support on Slack

• Enterprise

  Custom

  • Unlimited traces
  • Hosted, VPC or on-prem deployment
  • SSO and enterprise-grade RBAC
  • Real-time guardrails
  • Dedicated CSM
  • 24/7 support

Plans as listed on the official pricing page on 24 Oct 2026. Prices can change and may vary by country or exclude tax. Check the official pricing page.

Galileo alternatives

The main Galileo alternatives listed on Debuly are Braintrust, Langfuse, Fiddler AI, Maitai, Argilla and Arize AX. The table compares pricing and platforms; order is A–Z.

1. 

   Argilla

   Open-source data curation platform where AI engineers and domain experts collect human and AI feedback to improve LLM datasets.

   Developer platformsAutomation and agents
   Open sourceVisit Argilla (official site)
2. 

   Arize AX

   Observability and evaluation platform for AI agents and LLM apps, with tracing, issue detection and experiments on a free plan.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Arize AX (official site)
3. 

   Braintrust

   AI observability and evaluation platform for tracing LLM apps and agents, scoring outputs and catching regressions before release.

   Developer platformsCoding
   FreemiumVisit Braintrust (official site)
4. 

   Fiddler AI

   Control plane for enterprise AI agents: observability, evaluations, inline guardrails and governance for agentic and predictive models.

   Developer platformsAutomation and agents
   FreemiumVisit Fiddler AI (official site)
5. 

   Langfuse

   Open-source LLM observability and evaluation platform for tracing agents, managing prompts and scoring outputs.

   Developer platformsAutomation and agentsCoding
   FreemiumVisit Langfuse (official site)
6. 

   Maitai

   Control plane for production AI: monitor traffic, fine-tune models and version workflows and agents from live usage.

   Developer platformsAutomation and agents
   PaidVisit Maitai (official site)

Galileo FAQ

Sources & verification

Facts on this page come from the official sources below, each with the date we read it. Written by Debuly; the vendor did not review this page.

Image credits

• Logo: from framerusercontent.com · fetched 23 Oct 2026 · Vendor trademark, used only to identify the product; removed on request
• Cover: none collected yet, so a generated tile is shown.

Spot an error? Report an issue with this listing. We also remove a logo or image on the owner's request.