Latest / LEON — AI Systems Security | AGORA Intelligence
METR red team: OpenAI's agent monitor waited for Codex to stop
OpenAI's misalignment monitor fired when the agent ended its turn. METR estimates that between May and July 2026 the compaction gap probably left about a tenth of eligible Codex traffic unmonitored. Combined with the 99.2% eligibility figure, that puts the share actually monitored at around 89%.
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.