OpenAI announces GPT-5.3-Codex, an agentic coding model that combines frontier coding performance with general reasoning capabilities. The model achieves state-of-the-art results on SWE-Bench Pro and Terminal-Bench while being 25% faster than predecessors. It can handle long-running tasks including research, tool use,
•9m read time• From openai.com
Table of contents
Frontier agentic capabilitiesAn interactive collaboratorHow we used Codex to train and deploy GPT‑5.3-CodexSecuring the cyber frontierAvailability & detailsWhat’s nextAppendixSort: