Just ask and Codex can now spin up subagents
I tried OpenAI's newly launched subagents feature for Codex and it's awesome. I tried it via the Codex app on macOS, and the UI looks good as well.
Although spawning subagents consumes many more tokens than a single agent for the same task, it works faster because multiple agents are working on different tasks in parallel. By the way, the agent only spawns subagent(s) when you specifically ask it to. Codex docs have much more info about this and other related things.
I also learned from Simon's post that Codex also lets you define and use custom agents:
Codex also lets you define custom agents as TOML files in ~/.codex/agents/. These can have custom instructions and be assigned to use specific models - including gpt-5.3-codex-spark if you want some raw speed.
For your info, the subagents feature is only new to Codex as it's already available in Claude Code, Gemini, Cursor, OpenCode, etc.