OpenAI (ChatGPT / Codex)
Coding plans · reviewed by Stan K. · Source: data snapshot · data as of 2026-11-11 · scored at Plus
Codex rides the ChatGPT plans — and since the limits merged, your chat, Work and Excel share one opaque pool with your coding.
#8 of 10 in AI Coding Plans
Tiers — what each dollar buys
| Tier | $/mo | Annual | Quota & notes |
|---|---|---|---|
| Go | $8 | — | Codex "Limited" · may include ads · entry tier, the coding tier is Plus |
| Plus scored | $20 | — | Expanded Codex + Codex Cloud · merged ChatGPT+Codex limits |
| Pro | $100 | — | "From $100" — no fixed price on the pricing page anymore · longer Codex/Work sessions · monthly API credits (100/200/500 levels) |
Agent support: Codex clients only — ChatGPT desktop (Codex mode), Codex CLI, Codex IDE extension, Codex web — and the pool is shared with ChatGPT, Work, Excel and Workspace Agents.
PlanScore breakdown — why 60/100
10 plan criteria × 10 points; PlanScore = mean × 10. Low pillars are visible on purpose — that's the point of a score you can audit. Methodology →
What it does well
- ▸Codex limits compare well to Claude's walls — "codex usage limits are better than the Claude daily 5h limit" (HN 49337602)
- ▸GPT-5.6 Terra/Luna in Codex; GPT-6 reasoning on Plus; model menu keeps up with the frontier
- ▸Pro tier bundles monthly API credits (100/200/500 levels with eligibility)
- ▸Codex Cloud included from Plus
Where it costs you
- ▸Since Nov 2026 Codex and ChatGPT limits are merged — one pool for chat, coding, Work, Excel, Workspace Agents
- ▸No published numbers for the 5h/weekly windows; Support does no resets
- ▸Pro is "from $100" — the pricing page no longer shows a fixed price
- ▸Go $8 shows ads and Codex stays "Limited"
Hidden traps — the stuff the pricing page doesn't say
The Nov 2026 merge put Codex, ChatGPT, Work, Excel and Workspace Agents on one shared pool — "usage limits in codex are hit more often" after the change (HN 50320588). Your chat habit now eats your coding quota.
No published token/request numbers; the "banked" reset refreshes both the 5h and weekly window, and Support explicitly does not do manual resets.
The pricing page dropped the fixed $200 — actual Pro cost depends on configuration. Verify the live price before budgeting.
The $8 Go plan "may include ads" and keeps Codex at Limited — it is not a cheap coding ticket.
Know the walls before you pay: read the traps above, then decide with eyes open.
See OpenAI (ChatGPT / Codex) plansAffiliate link where available — never affects the score.
- official ChatGPT pricing (official) verified
- official OpenAI Help: using Codex with your ChatGPT plan verified
- community HN: 5h limit dies mid basic task on the $20 plan (2026-09) verified
- community HN: ChatGPT and Codex usage limits merging (2026-11) verified
- community HN: Pro single report — "essentially working 24/7" without limits (2026-05) verified
Data as of 2026-11-11. Prices and limits change monthly — report anything stale. How we verify: methodology.
Closest alternatives
~1.7B tokens/month for the $20-equivalent tier — but you can only buy it a year at a time, and the q
Kimi Code from the $19 tier, a rolling 5-hour limit with NO weekly cap — and an October 2026 rebuild
The multi-model aggregator plan: 6,000 requests/5h and 90,000/month across Qwen, GLM, Kimi and MiniM
OpenAI (ChatGPT / Codex) plan review
In-depth analysis · last updated 2026-11-11
Tiers and annual discounts
Go $8 (ads, Codex Limited), Plus $20 (expanded Codex + Codex Cloud), Pro "from $100" — the pricing page no longer commits to the old fixed number. No annual discount for individuals (annual billing is Business/Enterprise only). Scored at Plus, the tier people actually buy for Codex.
Worst value tier
Go $8: it may include ads and Codex stays "Limited" — you are paying for a chat tier, not a coding plan. The distance between Go and Plus is the whole product.
Hours before the wall (community)
- Plus 5h: dies mid-task — "can't even get through a basic sysadmin task on the $20 plan using Sol medium before I hit the 5hr one" (HN 49805509)
- Plus weekly: reachable in regular use — "sometimes I run into Codex's weekly limit and I need to purchase additional credits" (HN 47559143)
- Pro ($200-era, single report): "never even gotten close to token anxiety... essentially working 24/7" (HN 48311647)
- After the Nov 2026 merge: "usage limits in codex are hit more often" (HN 50320588)
Client lock-in
Codex clients only: ChatGPT desktop (Codex mode), Codex CLI, the IDE extension and Codex web — with the pool shared across ChatGPT, Work, Excel and Workspace Agents.
Versus Anthropic and OpenCode
Against Claude, users give Codex the friendlier wall: "codex usage limits are better than the Claude daily 5h limit" (HN 49337602). Against OpenCode Go the picture flips hard on value: for $10 OpenCode hands you a $60/mo model budget across 32 models in dollars you can actually read, where ChatGPT Plus gives you an opaque merged pool. If you already pay for ChatGPT, Codex is a strong free rider on that sub; as a standalone coding fuel plan it is mid-table.