OpenAI Releases GPT-5.2-Codex: Security-First Agentic Coding Model with Defensive Cyber Capabilities

OpenAI·medium signal

OpenAI launched GPT-5.2-Codex on April 21 with state-of-the-art performance on SWE-Bench Pro and Terminal-Bench 2.0. The model emphasizes cybersecurity as a first-class capability, with invite-only trusted access for defensive security professionals. Key improvements include context compaction for long-horizon work, stronger Windows support, and improved factuality. A smaller GPT-5.3-Codex-Spark variant delivers 1,000+ tokens/second for real-time coding.