FIX: Account for subscription benchmarks in quota estimates

[merefield]

What changed

• fold finalized subscription-funded Core, Extended, and DigBench usage into quota API-equivalent accounting
• defer both SPEND/NOW and 100%/FULL estimates while benchmark accounting is active or settling
• exclude API-key-funded benchmarks and fail closed when benchmark usage cannot be priced
• retain the existing hidden-session behavior while documenting and testing aggregate quota impact

Why

Benchmark sessions use ephemeral app-server threads. They consume subscription quota but do not enter the persisted rollout telemetry used by Codexometer, which could substantially understate both consumed and projected API-equivalent value.

Validation

• go vet ./...
• test -z "$(gofmt -l .)"
• go test -race -cover ./... -count=1
• go build -trimpath ./...

[Copilot]

🟡 Changes recommended

Two pricing paths can accept incomplete telemetry or incorrectly price cumulative long-context usage.

Once you've addressed the issues Copilot identified, you can request another Copilot review.

Pull request overview

Adds subscription-funded benchmark usage to quota estimates while deferring incomplete accounting.

Changes:

• Tracks and prices benchmark usage by billing source.
• Defers quota estimates during active/settling benchmarks.
• Adds documentation and accounting tests.

File summaries

File	Description
README.md	Documents benchmark quota behavior.
internal/ui/quota_api_eq.go	Defers API-equivalent display.
internal/ui/model.go	Integrates benchmark accounting lifecycle.
internal/ui/benchmark_quota.go	Implements benchmark usage accounting.
internal/ui/benchmark_quota_test.go	Tests accounting behavior.
internal/codex/benchmark.go	Exposes benchmark billing source.

Review details

• Files reviewed: 6/6 changed files
• Comments generated: 2
• Review effort level: Balanced

---

💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.