Heavy Codex user breaks down new usage limits: $500 plan burns 25% per feature
DulyDully · reddit · 2026-10-11
A year-long Codex user reviews OpenAI's usage rework and the value of each tier:
- The old $200/20x plan was effectively cheap unlimited frontier AI — unsustainable, likely a pre-IPO user-acquisition subsidy.
- $500 plan: ultra fast burns 1% per minute; a single end-to-end feature with debugging can consume 25% of quota. Only worth it for unlimited single Astra agent on fast.
- $100 plan: Astra burns it too fast, best for planning only, supports unlimited single 6.1 Sol agent; $200 roughly buys a few hours of daily Astra or unlimited dual 6.1 Sol agents.
- Complaint: resets that don't bank — topping up to 100% while also resetting the weekly timer makes the reset useless.
Verdict: $100/$200 plans are the best for most users; $500 only for heavy Astra needs.
More from coding & agent
- Polyphonic beta unifies Claude Code and Codex agents with cross-harness memory — RileyRalmuto · 2026-10-11
- Hamel Husain interviews benhylak on using simulations for agent evals — HamelHusain · 2026-10-11
- Anthropic's AI-Native SDLC Playbook: Versioned Intent, Skills for Policy, Hooks for Enforcement — bibryam · 2026-10-11
- Dev recreates employer Gitpod in 10 days, sparking debate on AI training data attribution — iamtrask · 2026-10-11
- Open-source tool reverse-engineers network traffic to turn apps into APIs for agents — Scary-Philosopher-77 · 2026-10-11
- AI sped up coding but broke QA: 64% of defects caught before prod in January — alex_verem · 2026-10-11