Skip to content

Add pricing v2 canonical convergence, website projection, and GPT-5.6 coverage#1

Merged
linqiang-max merged 10 commits into
mainfrom
feat/pricing-v2-canonical-convergence
Jul 10, 2026
Merged

Add pricing v2 canonical convergence, website projection, and GPT-5.6 coverage#1
linqiang-max merged 10 commits into
mainfrom
feat/pricing-v2-canonical-convergence

Conversation

@linqiang-max

Copy link
Copy Markdown
Contributor

Summary

Adds the Pricing V2 canonical convergence preview, website projection artifacts, and GPT-5.6 tier-aware OpenAI pricing coverage.

Major changes

  • Add Pricing V2 preview schema, generators, validators, tests, and evidence reports.
  • Add default-safe pricing and website projection artifacts.
  • Add GPT-5.6 Luna, Sol, and Terra model/API/history/snapshot coverage.
  • Add generic cache_write semantics for GPT-5.6 without asserting false duration-specific cache-write behavior.
  • Preserve truthful verifiedAt behavior and prevent regeneration-only timestamp refreshes.
  • Preserve public V1 compatibility while adding richer preview charge components.

Compatibility note

Public V1 changes are additive. cache_write appears as an additive pricing component; strict preview consumers should allow unknown/additional charge components.

Validation

  • Pricing V2 preview validation passed.
  • Focused Pricing V2 tests passed.
  • Website projection tests passed.
  • Exact feature HEAD and generated-artifact provenance were audited.
  • Snapshot/history integrity passed.
  • Existing non-GPT-5.6 pricing non-regression passed.

Known limitation

A prior broad dataset test run on Windows encountered filesystem permission/file-lock failures classified as environment-only. Pricing, schema, consistency, duplicate, source, and history integrity assertions relevant to this slice passed.

Deployment

No deployment performed.

AICostBudget added 10 commits July 6, 2026 23:29
Add OpenAI GPT-5.6 Sol, Terra, and Luna pricing coverage across canonical data, v2 preview artifacts, and generated public outputs.

Preserve GPT-5.6 cache writes as generic cache_write charges instead of duration-specific cache_write_5m semantics, and keep verifiedAt values truthful for existing projection rows.
Remove the hard-coded external Windows Website dataset path dependency.

Use a deterministic repository-local fixture plus configurable generator input.

Preserve projection semantics while making CI validation portable.
Replace root artifact upload with a narrow Pages artifact directory.

Use deterministic allowlisted staging for public dataset outputs.

Preserve documented /api/v1/* URLs while excluding repository internals.
@linqiang-max
linqiang-max merged commit 9e566fc into main Jul 10, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant