Codex is bundled into ChatGPT plans, with Pro now at three price points.
Codex (CLI, IDE, cloud, desktop) is included in ChatGPT Free through Enterprise and shares usage with ChatGPT Work. Plus has published local-message ranges per 5 hours; Pro comes at $100, $200 and $500 with no 5-hour limit.
Official pricing page ↗Values follow your controls on the Overview page (full-plan case).
$8/mo
GPT-6 Luna at Standard speed in the desktop app (subject to rollout).
Lightweight coding tier. Cheap model, so dollars stretch far in tokens but not in intelligence.
Best model: GPT-6 Luna (AA 38) · source
$20/mo
Local messages per 5h: Astra 5-45, GPT-6.1 Sol 15-160, Luna 350-3,000. Weekly limits may apply. Buy credits to extend.
OpenAI states message ranges, not dollars. Estimate assumes mostly GPT-6.1 Sol usage at typical agentic task sizes (2-15 credits per task).
Best model: GPT-6 Astra (AA 53) · source
$100/mo
Pro plans currently have no 5-hour limit; weekly limits apply. Higher multiples of Plus.
OpenAI does not publish the Pro multiples on the pricing page; estimate scales from Plus.
Best model: GPT-6 Astra (AA 53) · source
$200/mo
No 5-hour limit; weekly limits apply.
Estimate scales from Plus; see methodology.
Best model: GPT-6 Astra (AA 53) · source
$500/mo
Adds Ultrafast mode (burns included usage at 8x Standard for Astra / Sol).
Ultrafast consumes allowance faster, so tokens per dollar at Ultrafast are lower than the headline.
Best model: GPT-6 Astra (AA 53) · source
Dated reference points, a snapshot ledger you extend with npm run snapshot, and the plan changes that moved the numbers.
One snapshot so far. Each npm run snapshot adds a point.
One snapshot so far. Each npm run snapshot adds a point.
One snapshot so far. Each npm run snapshot adds a point.
One snapshot so far. Each npm run snapshot adds a point.