Reasoning to the top
Instant, Medium, High, Extra High and Pro on GPT-5.6 Sol and Sol Pro. Allowances are 5x or 20x Plus. At a cap, ChatGPT continues on another reasoning model until the reset.
Reviewed guide · September 1, 2026
Pro is sold in two sizes, $100 for 5x Plus usage and $200 for 20x. Both unlock Sol Pro reasoning. Text chat is unlimited everywhere; what Pro buys is headroom on the metered features. Here is what that headroom is and where it ends.
Reviewed: . Figures marked official come from OpenAI’s help center; everything else is labelled as observed or approximate.
Straight answer
Instant, Medium, High, Extra High and Pro on GPT-5.6 Sol and Sol Pro. Allowances are 5x or 20x Plus. At a cap, ChatGPT continues on another reasoning model until the reset.
Official: 40 files per project, 80 uploads per three hours, 512 MB per file, 2 million tokens per document, ~50 MB per spreadsheet, 20 MB per image.
The largest consumer allowances. Deep research resets 30 days after first use; images have a separate allowance.
Included with more headroom than Plus, still metered by tokens under OpenAI’s Codex rate card. Long autonomous runs are the usual way Pro users hit a wall.
Pick a tier
| Usage pattern | Tier | Why |
|---|---|---|
| Daily writing, analysis, some High reasoning | Plus $20 | Pro’s extra usage would go unused |
| High or Extra High reasoning most of the day | Pro $100 | 5x Plus covers heavy individual use; Sol Pro available |
| Codex agent runs for hours, daily | Pro $200 | Agent runs consume tokens fastest; 20x is the only personal tier sized for it |
| Team with shared data and admin needs | Business | Per-seat limits, workspace credits, governance; see the Business page |
Where Pro still stops
Unlimited text still has automated-use protection. Scripted or extreme volume gets paused regardless of plan.
512 MB, 2 million tokens, ~50 MB spreadsheets and 20 MB images apply to every plan including Pro.
A refusal is a content decision, not a quota. Upgrading changes nothing.
A single sprawling task can consume a large share of a day’s allowance. Scope it.
Alternatives
If a cap lands mid-task and you cannot wait for the reset, the practical fix is a second tool. None of these are affiliated with OpenAI; pick whichever fits the job.
Strong writing, analysis and long documents. Has its own five-hour and weekly limits.
AlternativeBest inside Google Workspace and for Google-context tasks.
ResearchResearch answers with web sources and citations.
AlternativeUseful inside Microsoft 365 and Windows.
Multi-modelSeveral models in one workspace so one provider’s cap does not end your day.
Disclosure: MultipleChat is built by the team that runs this site. It is listed as one option among several; nothing on this page is paid placement.
Official sources
Checked September 1, 2026. OpenAI changes plan rules without notice; these pages and your own in-app counter are authoritative.
Which reasoning options each plan gets, unlimited text on Free and Go, and what happens at a reasoning limit.
Current plan names, prices and what each tier includes.
Per-file size, files per project, rolling upload cap, Free daily cap and storage limits.
Token-based Codex usage, credits and plan coverage.
How deep research allowances are counted and reset.
FAQ
$100 per month for 5x Plus usage or $200 for 20x, both with Instant through Pro reasoning (GPT-5.6 Sol and Sol Pro), 40 files per project, the largest deep-research allowance and the most Codex headroom on a personal plan.
Text chat is unlimited on every plan. Pro raises the reasoning, image, upload, research and Codex allowances substantially but they are still finite, and abuse-prevention safeguards apply.
The highest reasoning option in GPT-5.6, available on Pro, Business and Enterprise. It is metered like other reasoning options.
Start at $100 unless you already exhaust it. $200 is four times the usage for twice the price only if you actually use it; there is no other feature difference.
Yes, Pro includes more Codex allowance than Plus, but Codex remains token-metered under OpenAI’s rate card, so long agent runs still consume it.
No. OpenAI states that Support does not reset ChatGPT or Codex usage limits.