Codex Now Scans Your Repos and Drafts Security Fixes

OpenAI updated Codex at DevDay 2026 with Security Cloud, which scans GitHub repositories on demand or on a schedule, investigates what it finds, filters out duplicates, and prepares fixes directly in the cloud. The same event brought a paid Ultrafast tier that speeds up token generation in Codex by up to eight times, reaching 300 tokens per second, but costs six times the standard rate: for GPT-6 Astra that is $60 per million input tokens and $300 per million output tokens. Ultrafast comes through the new $500-a-month Pro 500 subscription or an Enterprise plan.

Automated repo scanning with ready-made fixes is exactly what takes the routine part out of review, as long as the agent really filters duplicates instead of dumping false findings on the developer for the sake of a report. The Ultrafast math is simple: the price goes up six times for eight times the speed, and for me that only makes sense on tasks where an agent is genuinely stuck waiting on the model, not for ordinary background work. For most day-to-day agent tasks I would stay on the regular tier and not pay for speed I do not need.