Originally published on AI Cost Notes. Numbers checked 2026-10-30 against vendor docs.
Checked 2026-10-30 against vendor pages. Numbers move; re-check the same day you publish.
OpenAI's Help Center (updated ~Oct 29) says the planned Pro 200 cut is canceled. Pro 100 and Pro 500 now get 2× their previous Work + Codex allowance; every Pro 200 seat is on the higher Work + Codex allowance. Ultrafast stays Pro 500 only. Plus is unchanged.
That reset is the news. The $100–$200 shopping question underneath it is still: which sticker buys which meter?
What $100 / $200 buys (product + API)
| Plan | Sticker | Included product usage (this week) | Dollar API credits with the sub |
|---|---|---|---|
| ChatGPT Pro 100 | $100/mo | Pro chat; Work + Codex at the new 2× allowance; no Ultrafast | $100/mo after claim + link one API org † |
| ChatGPT Pro 200 | $200/mo | More than Pro 100; Work + Codex at the unified higher allowance (cut canceled); no Ultrafast | $200/mo † |
| Claude Max 5x | $100/mo | 5× Pro usage (rolling 5h + weekly) | $100/mo on Claude Platform ‡ |
| Claude Max 20x | $200/mo | 20× Pro | $200/mo ‡ |
| Cursor Plus | $8/mo | Base weekly Cursor-model pool | None; no third-party models |
| Cursor Premium | $30/mo | 4× Plus weekly pool | Third-party at API list from prepaid |
| Cursor Super | $100/mo | 15× Plus weekly pool | Third-party at API list from prepaid — no matching API grant |
| Cursor Ultra | $200/mo | 50× Plus weekly pool | Same prepaid rule |
† About ChatGPT Pro tiers · API credits benefit: Pro 100/200/500 → $100/$200/$500 API per billing period. Active ≥7 days, claim, link one org. Unused credits expire, no rollover. Credits fund the API Platform — they do not raise ChatGPT / signed-in Codex limits and do not cover Codex usage through API keys.
‡ claude.com/pricing compare table: Max 5x $100 / Max 20x $200 monthly API credits. API credits for Max/Team: since the Oct 7 Max/Team grant, credits cover Claude API, Managed Agents, Agent SDK, Console playground — not interactive Claude Code or extra usage in Claude apps. Also no rollover.
Cursor: Models & Pricing — Plus $8 / Premium $30 / Super $100 / Ultra $200. Weekly pool is for Cursor models (Grok/Composer). Claude/GPT/Gemini on Premium+ bill at API list from prepaid credit.
One number to remember
On ChatGPT Pro and Claude Max at these stickers, you still get roughly $1 of platform API credit per $1 of plan (after claim/link). That API bank does not refill interactive Codex (OpenAI) or interactive Claude Code (Anthropic). Cursor Super/Ultra spend the same $100/$200 on a bigger weekly Cursor-model pool, not on a free Claude/OpenAI API bank.
Pick by the meter you burn:
- API agents outside the chat app → Pro or Max win on the explicit grant.
- Work/Codex inside ChatGPT this week → Pro 100/200 both got the Work+Codex bump; Ultrafast only on Pro 500.
- Live in Cursor on Grok/Composer → Super/Ultra are the volume play; Claude inside Cursor still burns prepaid at list.
No affiliate links in this draft.
Top comments (0)