Version updated for https://github.com/Rul1an/llm-cost to version v1.3.0.

• This action is used across all versions by ? repositories.

Go to the GitHub Marketplace to find the latest changes.

Release notes

📣 Announcement: FinOps Validation Suite (v1.3.0)

“FinOps-as-Code” is now Enterprise Ready.

We are proud to announce the release of llm-cost v1.3.0, introducing the FinOps Validation Suite. This release transforms llm-cost from a developer tool into a trusted financial instrument, capable of audit-grade cost validation.

🎯 The Problem

Data Engineers and Platform Teams often struggle with:

• Drift: “Why is the bill 5% higher than the estimate?”
• Opacity: “Which models caused the variance?”
• Compliance: “Can we trust this data for chargeback?”

🛡️ The Solution: FinOps Validation Suite

We have implemented a rigorous, multi-tier testing infrastructure that runs on every change to the codebase.

1. P0 & P1 Certification (finops_p0, finops_p1)

Every commit is validated against strict FinOps criteria:

• ✅ Determinism: Zero-bit variance in output generation.
• ✅ Schema Integrity: Validates input/output against FOCUS v1.1 specs.
• ✅ PII Guard: Ensures no Personal Identifiable Information leaks into billing reports.
• ✅ Scale: Verified up to 1M rows and 2000+ unique cardinality.

2. FinOps Reports in CI/CD 📊

No more digging through console logs. Every Pull Request now generates a Cost Integrity Card directly in GitHub, reporting:

• Drift (BPS): Exact variance vs. reference.
• Audit Hash: SHA256 of the calculation logic.

3. Fail-Fast Policy

We have adopted a strict Fail-Fast policy for calibrate. Using corrupt or mismatched billing data will now trigger an immediate exit (Code 2), preventing “garbage-in/garbage-out” scenarios in your data pipeline.

🚀 Get Started

The suite is active by default. For Platform Engineers, integrate the reporting step:

- name: Generate Audit Report
  run: python3 tools/finops/render_report.py

Build trust in your AI spend.

• Rul1an
• GitHub Actions