On this page
For anyone already paying for ChatGPT and hitting limits—or weighing whether $20, $100, $200, or $500 a month is the right number—the choice is mostly about price, how fast the allowance burns, and which Pro-only tools actually matter. OpenAI's personal lineup now runs Free → Go → Plus → Pro, with Pro split into three usage tiers. The numbers below come from the ChatGPT pricing page and the Help Center articles for Plus and Pro tiers.
Monthly prices
ChatGPT Plus is $20 per month, billed monthly. The Help Center says annual billing and paying for multiple months in advance are not supported for Plus. API usage is not included and is billed separately.
ChatGPT Pro is sold as three monthly plans in US dollars:
| Plan | Monthly price (USD) | Ultrafast |
|---|---|---|
| Plus | $20 | Not on Plus |
| Pro 100 | $100 | Not included |
| Pro 200 | $200 | Not included |
| Pro 500 | $500 | Included |
Help Center wording is clear on the usage ladder: Pro 200 includes more usage than Pro 100, and Pro 500 offers the highest included usage of the three. Pro is also billed monthly; annual billing and multi-month prepay are not available.
Related reading: for API token bills rather than a ChatGPT subscription, see Claude API Pricing (October 2026).
Free, Go, Plus, and Pro at a glance
Before moving from Plus to Pro, it helps to see where both sit on the personal ladder. On the pricing page:
| Plan | Rough role | Headline price |
|---|---|---|
| Free | Everyday text chats | $0 |
| Go | Longer chats, more tools than Free | About $8/month in the US; localized in some markets (Introducing ChatGPT Go, January 16, 2026) |
| Plus | Advanced work and productivity | $20/month |
| Pro | Research and coding, three usage tiers | From $100/month (Pro 100 / 200 / 500) |
Go sits between Free and Plus: more messages with tools, more uploads and image creation, longer memory, and—like Free—it may include ads. Plus and Pro remain the ad-free personal paid tiers in OpenAI's consumer lineup as described on that Go launch post. Exact Go checkout prices can differ by country; the US list figure is about $8 a month.
The pricing feature table also shows how context steps up. For GPT Instant, Free lists a 27K total context window, Go and Plus list 54K, and Pro lists 128K. For GPT Reasoning, Go and Plus list 256K and Pro lists 400K. Instant input is described as about 12 pages on Free, about 40 on Go and Plus, and about 250 on Pro; Reasoning input is about 320 pages on Go and Plus and about 680 on Pro. Those page estimates are OpenAI's own approximations and can change with features and memory in use.
Standard, Fast, and Ultrafast—how allowance is spent
On Pro, speed modes change how quickly the plan's included allowance burns. Help Center rates for the same model:
- Fast uses included allowance at 2.5× the Standard rate.
- Ultrafast on GPT-6 Astra and GPT-6.1 Sol uses included allowance at 8× the Standard rate for that model, then draws from purchased credits at 6× the Standard credit rate after the included allowance is gone.
Those multipliers are consumption rates, not a promise that every reply will feel 2.5× or 8× faster. Usage still depends on the model, context size, reasoning, tools, retrieval, caching, and the speed mode in use.
Ultrafast is available only on Pro 500 among Pro plans. It is selected in the model picker. At launch, buying credits on Pro 100 or Pro 200 does not unlock Ultrafast.
Model allowances still exist on every paid plan. When a model hits its limit, it can be temporarily unavailable until the allowance resets; ChatGPT shows a reset time when it has one. Hitting a limit does not cancel the subscription. For eligible features, credits can be bought after included usage is gone—credits apply to some Codex and ChatGPT Work / Office surfaces, not to every model or feature.
Pro 200 grandfather through October 29, 2026
Pro 200 is open again for new subscriptions, but new sign-ups that are not eligible for grandfathering get a lower included usage allowance than the older Pro 200 offer, still at $200/month.
Grandfathering applies to an active Pro 200 subscription at any point from September 22, 2026 through 10 a.m. Pacific Time on September 29, 2026. Eligible customers can keep the previous included allowance through October 29, 2026 while the Pro 200 subscription stays active. After that date, the price stays $200/month, but included usage moves to the lower allowance. Keeping the older allowance does not upgrade the plan to Pro 500 and does not add Ultrafast.
If a Pro 200 subscription ended and the same account resubscribes, that September window is still required to qualify for grandfathering through October 29. Scheduling a cancellation alone does not end access until the current billing period finishes.
What Plus includes vs what Pro adds
Plus ($20) is pitched for advanced work: advanced reasoning models with GPT-6, expanded messages and uploads, stronger image creation, expanded deep research, expanded memory and context, projects / scheduled tasks / custom GPTs, expanded Codex usage, ChatGPT Work on desktop, web, and mobile, and early access to new features. Voice, file uploads, and image generation are part of the Plus feature set described on the Help Center and pricing pages. Limits still apply—especially under high demand.
Pro is pitched for research and coding on top of Plus. The pricing page highlights:
- A choice of three usage tiers (100 / 200 / 500)
- Pro reasoning powered by GPT-6 Astra
- Longer Codex and ChatGPT Work sessions
- Dot, the always-on agent (Pro only on the personal comparison table)
- Unlimited and faster image creation (subject to OpenAI's abuse guardrails)
- Maximum deep research and maximum memory / context
- Expanded projects, tasks, and custom GPTs
- Research previews of new features
- ChatGPT Space listed for Pro (not Free / Go / Plus on that table)
On models, Free and Go show no GPT-6.1 Sol / GPT-6 Astra access on the comparison grid, while Plus shows Yes and Pro shows Expanded. Response times are listed as limited on Free and Go, Fast on Plus, and Ultrafast (on eligible tiers) on Pro—which lines up with Ultrafast living on Pro 500. Codex is Limited on Free and Go, Yes on Plus, and Expanded on Pro.
Marketing copy on the Pro plan page still talks about “unlimited” messages and uploads; the Help Center is more precise: allowances vary by model and tier, and a model can go temporarily unavailable until reset. Treat “unlimited” as the product's everyday-chat framing under abuse guardrails, not as a hard guarantee that every Pro model is always open.
Who should stay on Plus—and who should pick which Pro
Plus fits when $20 covers a typical week: writing and editing, solid reasoning, Codex and ChatGPT Work at Plus levels, deep research without living in Astra Ultrafast, and models rarely lock until reset. Plus is still the default “serious personal” plan for most people who outgrew Free or Go.
Pro 100 ($100) fits when Pro-tier reasoning (Astra), longer Codex / Work sessions, Dot, larger context windows, and more included usage than Plus are needed—but Ultrafast is not required and Pro-level sessions are not burned through every day.
Pro 200 ($200) fits when Plus is already too tight and Pro 100's included usage would still feel small. New Pro 200 subscriptions without grandfathering get the lower included allowance; for accounts in the September window, the older allowance only lasts through October 29, 2026.
Pro 500 ($500) fits only when Ultrafast on GPT-6 Astra / GPT-6.1 Sol is specifically needed, or for heavy research / coding that will use the highest included allowance. Ultrafast burns included usage at 8× Standard, so it is a speed trade against the monthly pool—not free horsepower.
When the real bill is API traffic in apps under development, none of these ChatGPT subscriptions replace an API account. Plus and Pro Help Center pages both say API usage is billed separately. For how other products meter seats and credits, see also How Cursor Usage Limits Work on The X Subscription and Legacy Plans (October 2026) or Claude Code vs Codex: What $20 Gets You and Which to Pick.
A short decision checklist
- Need ad-free advanced models at the lowest paid step → Plus ($20).
- Need Dot, Pro Astra, bigger context, longer Codex/Work → start at Pro 100, step up if resets keep hitting.
- Need Ultrafast on Astra / Sol → Pro 500 only.
- Building products on tokens → use the API, not Plus/Pro included chat allowance.
- On grandfathered Pro 200 → note the October 29, 2026 cutoff for the older included pool.
Exact per-model message caps are not published as fixed public numbers on the Help Center pages used here, and they can move with demand. For day-to-day limits on an account, the in-product model picker and usage settings are the live source. For reference only—heavy weeks on a specific tier can be shared with the site owner through the contact link in the site footer.