Codexometer v0.11.1

Highlights

• Corrects quota API-equivalent estimates by incorporating finalized subscription-funded Core, Extended, and DigBench usage; estimates are deferred while benchmark accounting is incomplete.
• Fails closed on partial benchmark telemetry and aggregate long-context usage that cannot be priced accurately.
• Separates benchmark setup and model-turn deadlines so slow platform setup no longer consumes model execution time.

Full changelog: v0.11.0...v0.11.1

Codexometer v0.11.0

Adds verified cross-platform installers with checksum validation and atomic upgrades, publishes Linux, macOS, and Windows artifacts for AMD64 and ARM64, centralizes release version metadata, and removes duplicate pull-request CI runs.

Codexometer v0.10.0

Codexometer's Monitor now runs automatically with Pause/Resume and Reset controls, adaptive polling, bounded history, and keyboard session management.

The shared status light distinguishes healthy idle, active work, visible sessions needing attention, confirmed outages, and unknown or paused state.

Codexometer v0.9.6

Highlights

• Add the L keyboard shortcut for clearing accumulated Benchmark results, surfaced in the control label and safely disabled during active runs.

See PR #28 for implementation and validation details.

Full changelog: v0.9.5...v0.9.6

Codexometer v0.9.5

Highlights

• Preserve out-of-scope benchmark results across scoped runs, replace matching rows, and add explicit Clear All behavior while Run All starts clean.
• Refine Benchmark controls and Scope editing with transactional Cancel/Done actions and input-aware keyboard highlighting.
• Refresh Standard API pricing, including the promotional GPT-5.6 Sol rates retrieved on August 23, 2026.

See PR #27 for implementation and validation details.

Full changelog: v0.9.4...v0.9.5

Codexometer v0.9.4

Highlights

• Strengthen DigBench fidelity with continuous scoped play, compact model-facing state, dynamic creative-mode guidance, and a two-hour discovery allowance.
• Improve live DigBench reporting with exact published-condition labels, concise level progress, and full sanitized transcripts.

See PR #26 for implementation and validation details.

Full changelog: v0.9.3...v0.9.4

Codexometer v0.9.3

Highlights

• Remove the stray one-cell gap before the top-right corner of titled informational frames.
• Keep borders continuous when a title action such as Close is present, without changing action spacing or click targets.

See PR #25 for implementation and validation details.

Full changelog: v0.9.2...v0.9.3

Codexometer v0.9.2

Highlights

• Add a lower-right Copy control and c shortcut to export Benchmark results as a clean eight-column Markdown table.
• Include every completed or active row—including filtered and off-screen data—while retaining current sorting and rank weighting and excluding interactive controls.

See PR #24 for implementation and validation details.

Full changelog: v0.9.1...v0.9.2

Codexometer v0.9.1

Highlights

• Restore observed-run rankings for completed DigBench attempts, with rank sorting and Cost, Balanced, and Speed weighting.
• Keep DigBench rankings separate from Core and Extended; wins lead, losses/incomplete attempts follow, and stopped rows remain unranked.
• Document the server-assigned seed caveat clearly: individual DigBench rankings are exploratory rather than controlled model comparisons.

See PR #23 for the ranking implementation and validation details.

Full changelog: v0.9.0...v0.9.1

Codexometer v0.9.0

Highlights

• Add optional DigBench support, discovering all available games from /games at launch and hiding the suite when DIGBENCH_API_TOKEN is absent.
• Run selected DigBench games across scoped model/reasoning combinations with exact remote-session confirmation, live level/step progress, authoritative win detection, and stoppable runs.
• Organize built-in benchmarks into Codexometer Core and Codexometer Extended, with per-task scope selection plus clearer Run Scope and Run All behavior.
• Expand benchmark run details with a navigable, copyable workflow transcript covering safe prompts, tool requests/responses, moves, states, final responses, usage, and API-equivalent cost.
• Support isolated benchmark API-key authentication in preference to the prevailing Codex login when supplied, while keeping DigBench credentials out of spawned Codex environments.
• Harden privacy and reliability with benchmark-only bounded capture, runtime-identifier and terminal-control redaction, whole-run timeouts, safe interruption, and resilient remote-session handling.

See PR #21 for the complete implementation and validation details.

Full changelog: v0.8.0...v0.9.0