codex-cost-meter v1.2.6

• Estimate Auto-review safety checks at $0 from the October 6 documentation observation boundary, assuming ChatGPT sign-in; preserve historical estimates, token totals, and ordinary model rates.

codex-cost-meter v1.2.5

• Add dated GPT-6 Sol, GPT-6 Luna, and GPT-6.1 Sol pricing; restore GPT-5.4/5.5 and Pro long-context pricing to evidenced launch boundaries.
• Document the pricing refresh procedure, source evidence, and historical limitations.

codex-cost-meter v1.2.4

• Streamline human reports with compact identity, nested token components, scoped turn time, conditional model detail, and concise pricing qualifications while preserving full JSON provenance.

codex-cost-meter v1.2.3

• Add GPT-6 Astra pricing for Standard, Fast, cache writes, and long context from September 3; correct Sol's reduced-price start date to August 21.

codex-cost-meter v1.2.2

• Keep macOS Homebrew LaunchAgents on the stable opt executable path across upgrades.

codex-cost-meter v1.2.1

• Developer ID-sign and notarize macOS release binaries before packaging and publication.

codex-cost-meter v1.2.0

• Price codex-auto-review through its dated GPT-5.4 and GPT-5.6 Luna proxy history and expose the qualified mapping in human and structured reports.
• Condense human service-tier output, wrap pricing provenance, render intentional empty-model states, and erase stale progress-line text.

codex-cost-meter v1.1.0

• Exclude proven legacy-fork rollout replay from usage totals while preserving the first genuine child request.

codex-cost-meter v1.0.0

• Add inclusive local-calendar corpus and project reports with explicit date bounds, composable grouping, sparse or explicit empty buckets, and metadata-backed candidate pruning.
• Reuse an existing rollout cache read-only when write access is unavailable, with explicit warnings when the cache cannot be updated or created.

codex-cost-meter v0.9.0

• Attribute rollout usage to Standard, Fast, or unavailable service tiers, with current Fast and long-context pricing plus per-model tier detail.