Models & reasoning

🎯Goal. Pick the Codex model that fits the task — speed, cost, or maximum capability.

1. 1Start with gpt-5.5 — OpenAI's strongest for complex coding, computer use, and research workflows. It's the default recommendation.
2. 2Switch to gpt-5.4-mini for a faster, lower-cost option on lighter tasks or subagents. gpt-5.3-codex-spark (research preview, ChatGPT Pro) is tuned for near-instant, real-time iteration.
3. 3Change model with /model during a session, or start with the -m flag: codex -m gpt-5.5. In the IDE, use the model selector below the input box.

✓You'll see. A heavy refactor on gpt-5.5 versus a quick edit on gpt-5.4-mini — capability where you need it, speed where you don't.

💳Cost. Model choice is the main cost lever — the mini model is markedly cheaper for routine edits; save the frontier model for hard problems.

💡Takeaway. Default to gpt-5.5, drop to gpt-5.4-mini for light or fast work, and switch any time with /model or -m.