ChatGPT Pro 100, 200 or 500: which AI plan fits? (updated limits)

October 30, 2026 update: OpenAI canceled the Pro 200 cut on October 29 and doubled the previous Work and Codex allowance for Pro 100 and 500. Business Premium receives more usage without an official figure for the increase. See the new ChatGPT Pro vs Business guide. The earlier version described a change that no longer applies.

In 60 seconds

If you pay for a premium AI plan, identify which work gets interrupted before switching subscriptions. The Pro 200 cut was canceled: the lower allowance increases to match the higher one, which existing higher-allowance subscribers keep. Retain a weekly record of tasks, limits and additional spending to decide using current limits.

A founder preparing proposals, a developer working with an assistant and an agent handling product requests have different requirements. You can combine a personal license, team seats and API usage, but each expense needs a specific purpose.

The change planned for October 30 was canceled

The official Pro help page and October 29 release notes, checked on October 30, 2026, supersede the earlier announcement. All Pro 200 subscriptions now have the same Work and Codex allowance for USD 200/month. New subscriptions include it; existing subscriptions update automatically without waiting for renewal or canceling and resubscribing.

Earlier announcement	Status checked on October 30, 2026
Pro 200 cut after temporary grandfathering	Canceled; the higher allowance stays and the lower one doubles to match. ChatGPT release notes.
Weekly GPT-6 Pro limits in Chat	Pro 200 aligns the lower limit with the higher; Pro 100 and 500 keep their limits. These are separate from Work/Codex. Pro help.
Email figures relative to Plus	Not used as current limits: OpenAI does not publish these multiples in the help page checked.
September email’s 62,500-credit grant	Its status after cancellation is unconfirmed / check your account. Do not budget it as a guaranteed benefit.
Other plans	Pro 100 and 500 double Work/Codex; Business Premium receives more usage without a quantified increase; Plus is unchanged and Standard is absent from the announcement. Business notes.

The official Work/Codex pricing documentation, checked October 30, says Pro plans currently have no five-hour limit. Weekly and model-specific limits still apply. No end date or guarantee of permanence is published for the current allowance.

Individual plans: verified prices and limits

ChatGPT Pro verified on October 30, 2026; other plans retain their October 28 verification. References are in USD per month for web subscriptions; the table uses US list prices where regional pages exist. This is not a final LATAM quote: confirm taxes, currency, purchase channel and availability at checkout. Each row links to an official source.

Plan	Reference price	Included usage and what to check
ChatGPT Plus	USD 20/month	Plus help. Limits depend on model and tool; API billed separately.
ChatGPT Pro 100	USD 100/month	Pro help, October 30: twice its previous Work/Codex allowance; no fixed completed-task count.
ChatGPT Pro 200	USD 200/month	Official release notes, October 30: cut canceled and allowances aligned. Check your panel.
ChatGPT Pro 500	USD 500/month	Pro help, October 30: twice its previous Work/Codex allowance; the only Pro tier with Ultrafast.
Claude Pro	USD 20 monthly; website displays USD 17/month with USD 200 billed upfront annually	Claude pricing. Compare annual commitment separately from flexible monthly billing.
Claude Max 5×	USD 100/month	Max help. 5× Pro per five-hour session; weekly cap also applies.
Claude Max 20×	USD 200/month	Max help. 20× Pro per five-hour session; weekly cap also applies.
Google AI Pro	USD 19.99/month, US	Google US subscriptions. Check the limit for the feature you use.
Google AI Ultra 5×	USD 99.99/month, US	Google US subscriptions. 5× AI Pro’s limits.
Google AI Ultra 20×	USD 199.99/month, US	Google US subscriptions. 20× AI Pro’s limits.

Anthropic does not promise a fixed message count: usage depends on conversation, model and features (official FAQ, October 28). Max adds USD 100/month in API credits on 5× and USD 200/month on 20×. You must claim them by linking a Console organization; new subscribers need seven days on the plan. These credits do not cover extra usage in Claude or interactive Claude Code. Check expiry and scope in the API credit help page, reviewed October 28.

Google lists Argentina, Brazil, Chile, Colombia, Mexico and Peru among AI Pro and Ultra supported locations, checked October 28. We also checked the Argentina, Colombia and Mexico pages: regional terms apply. Final tax-inclusive price and each feature’s availability for your account: verify. Do not extrapolate US pricing across LATAM.

Decision tree: which AI plan fits your work?

1. Does AI execute work inside your product or handle requests automatically? Evaluate API usage with a budget, consumption logs and an operational owner. Include infrastructure and review; see the real cost of an AI agent.
2. Do several people need company access and controls? Compare Business/Team seats and your existing suite’s option. An individual license price does not settle permissions, onboarding, offboarding or ownership of work. Use our ChatGPT, Claude, Gemini and Copilot business comparison for that decision.
3. Are you working individually, with limits interrupting important tasks? Test an intensive individual plan: Pro 100/200/500, Claude Max or Google AI Ultra. Choose using the task and measured interruptions. For more on Anthropic, read which Claude subscription to choose.
4. Is usage occasional, without blocking your work? Start with the smaller plan that includes your features: Plus, Claude Pro or Google AI Pro. Keep your current plan if it already handles the work.

For ChatGPT Pro vs Claude Max, the multiples use different baselines and windows. A subscription multiple cannot establish how many tasks another subscription will complete. Neither translates into tasks, billable hours or correct results. Our comparisons hub covers other decisions by need.

Measure a week before upgrading

Log real work using test data or authorized material. The official Work/Codex documentation, reviewed October 28, explains that model, context, reasoning, tools and speed mode affect consumption. Counting messages helps with recordkeeping, but cannot establish capacity on its own.

What to record each day	Why it helps
Task and acceptance criterion	Distinguish a draft from a usable result.
Plan, model, mode and tool	Compare the same configuration where possible.
Chat messages and Work/Codex sessions separately	Identify which usage causes interruptions.
Visible consumption before/after and reset time	Identify the specific limit; mark missing data as unavailable.
Blocks, waiting, retries and human review	Determine whether more allowance would solve the observed problem.
Cost of repeating a sample through the API	Compare measured billing with the license and its extras.

Repeat a representative task with the same material and acceptance criterion in an alternative. For an API trial, set a small budget and record input, output, caching, tools and retries according to that provider’s billing. Do not reconstruct an API cost from a subscription multiple.

October 29, 2026 update: for an OpenAI pilot, see how to claim ChatGPT Business or Pro API credits, including requirements, expiry and unconfirmed country availability.

Compare subscription + additional purchases + review against API consumption net of applicable credits + infrastructure + operations + review for an equivalent workload. Count credits only if you can use them for that workload before expiry. They are not cash or automatic savings. If the alternative needs extensive corrections, include that work even when the call costs less.

What to do with the updated limits

Check your plan, balance and renewal; save a usage-dashboard reference without private data. If you do not yet have a week’s record, start with the updated allowance and repeat the same tasks. You do not need to cancel and resubscribe. If the displayed usage differs from official terms, retain the notice and contact support.

Separate subscription, purchased credits and API consumption on your invoices; check currency, taxes, renewal and mobile-store charges. Review any promotional balance’s expiry. Avoid an impulsive annual commitment; test a real task first and read cancellation terms. Our Pro vs Business guide explains how to check account or workspace usage and when to mix Standard and Premium seats.

For a team building B2B agents in LATAM, keep instructions, reference documents, tools, permissions and acceptance criteria outside one provider’s exclusive chat history. Retain an evaluation sample and a manual fallback procedure. Before switching models, recheck quality, data access and error handling. Plans can change with limited notice; a documented workflow lets you evaluate another option without rebuilding the process from scratch.

When to wait, and the next step

If you do not know which limit you reach, which data you can send or who will review the result, settle those questions before buying more capacity. A more expensive plan cannot fix an ownerless workflow or establish that an agent is ready for production.

Bring one task, its acceptance criterion and your usage log to an AI diagnostic with Kiia. We can review which part fits a license and which needs an integration with a budget and monitoring.

ChatGPT Pro terms updated on October 30, 2026. Other plans retain their October 28 verification. Review planned every three months, or sooner after an announced provider change.