how to route your codex / GPT models for marketing
use GPT 6 Astra for thinking, judgment and planning, then let Sol coordinate the work and route tasks downstream
give Astra your campaign idea and goal / objective, then point it to the context you have
→ Astra · high reasoning
works with you to think through how the campaign should be done, weigh different approaches and shape the direction. use it with Matt Pocock's /wayfinder skill to map the path to completion, with tickets, evals to assess the work and decisions you need to be involved in
→ Sol · medium reasoning
turns the agreed plan into execution tasks, coordinates their order and gives each model the context it needs. if campaign copy depends on an approved offer, Sol waits for that decision before assigning the draft
→ Luna · high reasoning
handles execution that needs extensive context and precision, like writing a landing page from customer interviews, product documentation and your brand guidelines
→ Terra · medium reasoning
handles routine execution, like formatting approved copy for different channels, updating campaign records and applying specific edits
Sol evaluates the outputs against the requirements and requests revisions. decisions that need further judgment go back to Astra before your sign-off
Codex tip: A cost-efficient Luna + Sol agent tree, orchestrated by Astra.
Effort levels chosen by weighing DeepSWE’s pass rates, average cost per task, and agent steps. Hand this to Codex to set it up 👇