Literature Review Integrity Runs

Paired quality and integrity evaluation. Quality scores and integrity scores are separate measurements; integrity status is not a quality pass label.

No isolated integrity run selected.
Runs are compared within one release, on the same case set.

Capability Matrix

Agent
Claude
Codex
Gemini
Peak

Agent Capability Signature

Historical Evolution

Test Case Breakdown

Historical broad benchmark runs are quality-only records — not integrity evaluated.