ChatGPT Plus vs Pro: Complete Comparison and Report on Price Tiers, Usage Limits, Context Window, and API Credits
ChatGPT Plus and ChatGPT Pro are the two paid individual subscriptions OpenAI sells above the free and $8 Go plans. Plus costs $20 a month and has been the standard personal upgrade since 2023. Pro launched as a single $200 plan, gained a $100 tier on April 9, 2026, and is now sold in three tiers at $100, $200 and $500 a month. Both plans run the same GPT-6 generation of models in Chat, ChatGPT Work and Codex, use the same apps and share most features, so the decision comes down to how much usage, context and speed a subscriber needs, and whether they also pay for the OpenAI API.
The facts on OpenAI's own pages point to a clear split. Plus is the right plan for most individuals: it includes GPT-6.1 Sol, GPT-6 Astra, GPT-6 Sol and GPT-6 Luna, deep research, projects, Codex and ChatGPT Work, with published five-hour message ranges for coding work and the option to buy extra credits. Pro is worth its price for three kinds of user: people who regularly hit the Plus limits in Codex or ChatGPT Work, because Pro has no five-hour limit; people who need longer inputs, because Pro raises the reasoning context window from 256K to 400K tokens and the Instant window from 54K to 128K; and developers who already spend at least $80 a month on the OpenAI API, because every Pro tier returns its full monthly price as API credits that Plus does not get. Pro alone adds dots (OpenAI's always-on agent), GPT-5.6 Sol Pro, Composer predictions in supported regions, and Ultrafast mode on the $500 tier.
The $200 tier is the hardest to justify on price: it costs twice as much as Pro $100 and returns $100 more in API credits, so it pays off only for subscribers who need its higher usage allowance, which OpenAI no longer states as a multiple on its current pricing pages. Prices and plan details in this comparison were verified on OpenAI's official pricing pages, Codex and ChatGPT Work pricing documentation, plan pages and developer documentation on November 8, 2026.
··········
CHATGPT PLUS AND CHATGPT PRO: HOW THE TWO PLANS REACHED THEIR CURRENT SHAPE.
Plus has kept the same $20 monthly price since it launched, while the features behind that price have changed with every model generation. Today OpenAI's Plus plan page lists "Advanced reasoning models with GPT-6", expanded messages and uploads, more complex and accurate image creation, expanded deep research, expanded memory and context, projects, scheduled tasks and custom GPTs, expanded Codex usage, expanded access to ChatGPT Work on desktop, web and mobile, and early access to new features. The page adds that limits apply, without giving numbers.
Pro has changed much more in 2026. It started as one $200 plan. On April 9, 2026, OpenAI announced a $100 Pro tier in its developer forum's announcements category, describing it as offering five times the Codex usage of Plus, with all Pro features in ChatGPT, including the exclusive Pro model and unlimited Instant and Thinking models. A $500 tier followed and became the only individual plan with Ultrafast mode: OpenAI's Codex changelog lists Ultrafast on Pro $500 for GPT-6 Astra in the week of September 28 to October 2 and for GPT-6.1 Sol on October 8.
The most recent change came on October 29, 2026, when OpenAI doubled ChatGPT Work and Codex usage limits on all three Pro tiers and on Business Premium, announced by Thibault Sottiaux, who leads OpenAI's core products and platform work, in a post on his X account. Plus was not part of that change. The same announcement dropped a planned reduction of the Pro 200 allowance that subscribers had reported in the days before, as detailed in our report on how OpenAI doubled ChatGPT Pro usage limits and cancelled the Pro 200 cut.
........
Date | Plan event | Plan affected |
April 9, 2026 | $100 Pro tier announced, 5x Plus Codex usage at launch | Pro |
Sept 28 to Oct 2, 2026 | Ultrafast for GPT-6 Astra in Codex and Work | Pro $500 only |
October 8, 2026 | Ultrafast for GPT-6.1 Sol | Pro $500 only |
October 26 to 30, 2026 | GPT-6.1 Terra added to ChatGPT Work and Codex | Plus and Pro |
October 26 to 30, 2026 | Dot texting beta via iMessage and RCS in the US | Pro only |
October 29, 2026 | Work and Codex usage limits doubled | Pro tiers (not Plus) |
........
··········
CHATGPT PLUS VS PRO PRICING: $20 AGAINST THREE TIERS FROM $100 TO $500.
OpenAI's Codex and ChatGPT Work pricing documentation lists Plus at $20 a month and Pro "From $100/month", with plans at $100, $200 or $500 USD per month. The consumer pricing page at chatgpt.com shows Pro as "Your choice of 3 usage tiers" and names the tiers Pro 100, Pro 200 and Pro 500 in its API credit row. The pricing page FAQ says OpenAI offers monthly plans for Go, Plus and Business and annual plans for Business and Enterprise, so neither Plus nor Pro has an annual option with a discount.
The annual totals below are simply twelve monthly payments, computed in Python. The gap is large: a year of Pro 100 costs $960 more than a year of Plus, and a year of Pro 500 costs $5,760 more.
........
Plan | Monthly price (USD) | Cost over 12 months | Annual billing option | Monthly API credits |
ChatGPT Go | $8 | $96 | No | None |
ChatGPT Plus | $20 | $240 | No | None |
ChatGPT Pro 100 | $100 | $1,200 | Not listed | $100 |
ChatGPT Pro 200 | $200 | $2,400 | Not listed | $200 |
ChatGPT Pro 500 | $500 | $6,000 | Not listed | $500 |
........
Billing terms for Plus are set out in OpenAI's help article on the plan: payment is by credit card, the subscription can be cancelled at any time from the account settings, access continues until the end of the paid period, and cancellation must happen at least 24 hours before the billing date to avoid the next charge. Subscription fees are non-refundable, with a 14-day refund window for customers in the EU, the UK and Turkey. The ChatGPT API is billed separately from the subscription. For the full plan ladder from Free to Enterprise, see our guide to ChatGPT pricing for Free, Go, Plus, Pro, Business and Enterprise.
··········
MODELS IN CHATGPT PLUS AND CHATGPT PRO: SAME FAMILY, DIFFERENT ALLOWANCES.
Both plans run the current GPT-6 models. OpenAI's pricing table marks GPT-6.1 Sol, GPT-6 Astra, GPT-6 Sol and GPT-6 Luna as included on Plus and "Expanded" on Pro. The older GPT-5.6 Sol, GPT-5.6 Terra and GPT-5.6 Luna are included on Plus and "Unlimited", subject to abuse guardrails, on Pro. GPT-5 Thinking Mini is "Expanded" on Plus and unlimited on Pro. The only model OpenAI lists as Pro-exclusive in Chat is GPT-5.6 Sol Pro, and the Pro plan page describes "Pro reasoning powered by GPT-6 Astra".
OpenAI's developer guide describes the roles of the current models: GPT-6 Astra offers the highest intelligence, GPT-6.1 Sol balances speed, cost and intelligence with "near-Astra performance for complex work at a lower cost", GPT-6.1 Terra is fast and efficient and suited to visual reasoning, computer use and writing, and GPT-6 Luna is the fastest and most cost-effective. The GPT-6.1 Sol system card addendum, dated September 29, 2026, describes capabilities comparable to GPT-6 Astra with better speed and affordability, and classifies the model as Critical in cybersecurity and High in biological and chemical capability under OpenAI's Preparedness Framework. GPT-6.1 Terra became available in ChatGPT Work and Codex on Plus, Pro, Business, Enterprise and Edu in the week of October 26.
........
Model in ChatGPT | Plus | Pro |
GPT-6.1 Sol | Yes | Expanded |
GPT-6 Astra | Yes | Expanded |
GPT-6 Sol | Yes | Expanded |
GPT-6 Luna | Yes | Expanded |
GPT-5.6 Sol | Yes | Unlimited (abuse guardrails) |
GPT-5.6 Sol Pro | No | Yes |
GPT-5.6 Terra | Yes | Unlimited (abuse guardrails) |
GPT-5.6 Luna | Yes | Unlimited (abuse guardrails) |
GPT-5 Thinking Mini | Expanded | Unlimited (abuse guardrails) |
Legacy models | Yes | Yes |
........
The model list explains why many subscribers never need Pro. A Plus user can pick GPT-6 Astra or GPT-6.1 Sol for difficult work and GPT-6 Luna for quick tasks, the same models a Pro user chooses from. What changes is how often they can use the heavier models before reaching a limit, and the size of the inputs those models accept.
··········
CHATGPT PLUS VS PRO USAGE LIMITS IN CODEX AND CHATGPT WORK.
The usage gap is the main reason to move from Plus to Pro, and it is also where OpenAI publishes the least detail. For Plus, the Codex and ChatGPT Work pricing documentation gives ranges of local messages per five-hour period for each model, and warns that weekly limits may also apply and that cloud tasks can use more of the allowance than local messages. Local messages and cloud chats share the same allowance. OpenAI presents the ranges as estimates, because the real count depends on the length of the task, the size of the codebase and the reasoning effort.
For Pro, the same documentation states that "Pro plans currently have no five-hour limit" and directs subscribers to the usage dashboard for current limits. Weekly limits may still apply. OpenAI's April announcement described Pro 100 as offering five times the Codex usage of Plus, and the Pro 200 plan has been described by OpenAI as having 20 times higher limits than Plus. After the October 29 doubling, OpenAI did not publish new multiples, and the current chatgpt.com pricing page describes Pro only as "Your choice of 3 usage tiers", so the exact allowance of each tier today is not public.
........
Model (Codex and ChatGPT Work) | Plus: local messages per 5 hours | Pro 100, 200 and 500 |
GPT-6 Astra | 5 to 45 | No five-hour limit; weekly limits may apply |
GPT-6.1 Sol | 15 to 160 | No five-hour limit; weekly limits may apply |
GPT-6.1 Terra | 35 to 285 | No five-hour limit; weekly limits may apply |
GPT-6 Sol | 15 to 150 | No five-hour limit; weekly limits may apply |
GPT-6 Luna | 350 to 3,000 | No five-hour limit; weekly limits may apply |
........
Image generation also draws on these allowances: OpenAI's documentation says that on Plus it uses limits about three to five times faster on average, while the Pro plan card promises "Unlimited and faster image creation". Desktop voice in Codex costs $0.05 per minute from the existing usage budget on both plans. In Chat, everyday text conversations and chat history are marked unlimited on both plans, subject to abuse guardrails, and voice is "Expanded" on Plus and unlimited on Pro.
From December, OpenAI plans to move ChatGPT and Codex onto a single token-based usage limit, according to its help-center guidance, which we covered in our report on the shared ChatGPT and Codex usage limit. The help-center article could not be retrieved for this comparison, so the five-hour ranges above reflect the pricing documentation as published on November 8.
··········
CHATGPT PLUS AND PRO CONTEXT WINDOW AND INPUT SIZE.
Context is the clearest technical difference between the two plans, and OpenAI publishes it as fixed numbers. On Plus, GPT Instant has a 54K-token total context window with an input maximum of about 40 pages of text, and GPT Reasoning has a 256K-token window with about 320 pages of input. On Pro, GPT Instant rises to 128K tokens and about 250 pages, and GPT Reasoning rises to 400K tokens and about 680 pages.
........
Context limit in ChatGPT | Plus | Pro | Pro advantage |
GPT Instant total context window | 54K tokens | 128K tokens | 2.4x |
GPT Instant input maximum | About 40 pages | About 250 pages | 6.3x |
GPT Reasoning total context window | 256K tokens | 400K tokens | 1.6x |
GPT Reasoning input maximum | About 320 pages | About 680 pages | 2.1x |
........
The input maximum matters more than the window for people who paste long documents. A 300-page contract or a long research report fits within Plus's reasoning limit of about 320 pages, but a 500-page filing or a set of several long reports needs Pro's limit of about 680 pages to go in one message. The Instant gap is wider still: Plus accepts about 40 pages in a fast non-reasoning reply, while Pro accepts about 250. Both plans support file uploads, projects, data analysis and vision; the pricing table lists file uploads as available on both, while the plan cards describe uploads as "Expanded" on Plus and unlimited on Pro.
··········
[[ADPLUS_300x250_1]]
SPEED MODES AND CREDITS ON CHATGPT PLUS VS PRO.
OpenAI sells two faster generation modes in Codex and ChatGPT Work, and they consume usage at different rates. Fast mode is available on both plans and uses 2.5 times the included subscription usage, or twice the rate when paid with purchased credits. Ultrafast mode, available for GPT-6 Astra and GPT-6.1 Sol, uses eight times the included usage or six times the purchased-credit rate, and among individual plans it is limited to Pro 500. The pricing table lists Plus response times as "Fast" and Pro as "Ultrafast (on eligible tiers)", which means Pro 100 and Pro 200 subscribers do not get Ultrafast.
Plus and Pro subscribers who reach their limits can buy ChatGPT credits and continue without changing plan. OpenAI does not publish a single price per credit, saying that credit purchase prices and discounts depend on the plan or agreement, but it does publish how many credits each model consumes per million tokens.
........
Model | Input (credits per 1M tokens) | Cached input | Output | Fast mode with credits | Ultrafast with credits |
GPT-6.1 Sol | 50 | 2.5 | 250 | 2x | 6x (Pro 500 only) |
GPT-6.1 Terra | 25 | 1.25 | 125 | 2x | Not offered |
GPT-6 Luna | 2.5 | 0.25 | 12.5 | 2x | Not offered |
GPT-5.6 Sol | 100 | 10 | 500 | 2x | Not offered |
........
To show what these rates mean, consider a single coding task that sends 200,000 input tokens, of which 150,000 are served from cache, and returns 20,000 output tokens. Computed in Python with OpenAI's published rates, the task costs about 7.9 credits on GPT-6.1 Sol at standard speed and 15.75 credits in Fast mode, about 3.9 credits on GPT-6.1 Terra, and about 0.41 credits on GPT-6 Luna. OpenAI's own guide says a typical GPT-5.6 Sol task may use 2 to 15 credits. Because the dollar price of a credit is not published for individual plans, these figures compare the relative cost of models rather than a cash amount.
# Credits for one task: 200,000 input tokens (150,000 cached), 20,000 output tokens
rates = {"GPT-6.1 Sol": (50, 2.5, 250), "GPT-6.1 Terra": (25, 1.25, 125), "GPT-6 Luna": (2.5, 0.25, 12.5)}
uncached, cached, out = 50_000, 150_000, 20_000
for model, (inp, cache, outp) in rates.items():
credits = uncached / 1e6 * inp + cached / 1e6 * cache + out / 1e6 * outp
print(model, round(credits, 3))
# GPT-6.1 Sol 7.875 | GPT-6.1 Terra 3.938 | GPT-6 Luna 0.412··········
FEATURES ONLY CHATGPT PRO INCLUDES.
Most features are identical on the two plans. OpenAI's comparison table lists ChatGPT Work on desktop, web and mobile, plugins, Health, Finances, skills, search, deep research and apps for deep research, projects, scheduled tasks, Sites, the built-in browser, data analysis, vision, image generation with Thinking, study mode, memory sources, developer mode (beta) and voice with video for both Plus and Pro. Memory and memory with past chats are "Expanded" on both, although the Pro plan card promises "Maximum memory and context".
The differences that remain are concentrated in agents, sharing and research depth. Dots, which OpenAI calls "your always-on agent", are a Pro-only feature that keeps working between conversations and brings back results for review; in the week of October 26 OpenAI began rolling out a texting beta that lets US Pro users reach their dot through iMessage or RCS on Android, with text, photos and voice messages. ChatGPT Space and sharing plugins with a workspace are also marked Pro-only, and Composer predictions (beta) appear only for Pro, with regional limits.
........
Feature | Plus | Pro |
Dots (always-on agent) | No | Yes |
Dot texting via iMessage and RCS | No | Beta, US |
ChatGPT Space | No | Yes |
Share plugins with your workspace | No | Yes |
Composer predictions (beta) | No | Yes, regional limits |
Deep research | Expanded | Maximum |
Image creation | More complex and accurate | Unlimited and faster |
Codex | Yes | Expanded |
Excel, Word, PowerPoint and Google Sheets extensions | Yes | Expanded |
Voice | Expanded | Unlimited (abuse guardrails) |
Ultrafast mode | No | Pro 500 only |
Monthly API credits | No | $100, $200 or $500 |
Content used to train models | Opt-out available | Opt-out available |
........
Neither plan includes Business features such as SAML single sign-on or an admin console, and OpenAI's table lists the same setting for both on model training: "Opt-out available". OpenAI's Codex documentation also shows Codex Security for connected GitHub repositories as unavailable on both Plus and Pro, so neither individual plan replaces an Enterprise or Edu contract for that tool. Subscribers weighing Pro against a team plan can compare it in our ChatGPT Pro vs Business comparison.
··········
CHATGPT PRO API CREDITS: WHAT PLUS SUBSCRIBERS DO NOT GET.
The API credit benefit is the one Pro feature that can be converted into a cash equivalent. OpenAI's developer documentation lists monthly grants of $100 for the $100 tier, $200 for the $200 tier and $500 for the $500 tier, and names Plus, Go and Free as ineligible plans. The subscription must be paid and at least seven days old, free trials and complimentary subscriptions do not qualify, and the ChatGPT account and the API organization must use the same email address.
The credits come with restrictions that limit their value. Unused credits expire at the end of each billing cycle and do not roll over, and the first grant is prorated for the remainder of the current cycle. They cover API products such as Responses, Agents, Realtime, Images and built-in tools like web search, file search and Code Interpreter, but Codex usage through API keys is not supported, and the credits do not raise ChatGPT or signed-in Codex usage limits. Free credits are consumed before purchased ones, and when none remain, requests fail with an HTTP 429 error and the code credit_balance_exhausted. Downgrading to an ineligible plan such as Plus removes the organization link at the end of the billing cycle.
........
API credit rule | ChatGPT Plus | ChatGPT Pro |
Monthly grant | Not eligible | $100, $200 or $500 by tier |
Eligibility | Not applicable | Paid, at least 7 days old, same email on both accounts |
Rollover | Not applicable | None; expires at end of billing cycle |
First grant | Not applicable | Prorated for the current cycle |
Covers | Not applicable | Responses, Agents, Realtime, Images, built-in tools |
Excluded | Not applicable | Codex via API keys; ChatGPT usage limits |
How to claim | Not applicable | Settings, then Billing, then API credits benefit |
........
··········
WHAT CHATGPT PLUS AND PRO COST FOR THREE KINDS OF SUBSCRIBER.
Because Pro returns its price as API credits, its net cost depends on whether the subscriber already pays for the OpenAI API. The table below, computed in Python, adds each plan's subscription to the API spending left after credits, for monthly API budgets from zero to $400. It assumes the credits are fully used on eligible API products within the month, which is the only case in which they are worth their face value.
........
Monthly OpenAI API spend | Plus + API | Pro 100 + remaining API | Pro 200 + remaining API | Pro 500 + remaining API |
$0 | $20 | $100 | $200 | $500 |
$50 | $70 | $100 | $200 | $500 |
$100 | $120 | $100 | $200 | $500 |
$200 | $220 | $200 | $200 | $500 |
$400 | $420 | $400 | $400 | $500 |
........
Three conclusions follow from the arithmetic. A subscriber who does not use the API pays $80 a month more for Pro 100 than for Plus, so the upgrade has to be justified by usage, context or Pro-only features. A developer spending $80 or more a month on the API breaks even on Pro 100, and above that figure Pro 100 is cheaper than Plus plus the same API bill while also removing the five-hour limit. Pro 200 never costs less than Pro 100 on API spending alone: at $200 of API use the two are level, and below that Pro 100 is cheaper, so the $200 tier is a purchase of extra ChatGPT and Codex usage. Pro 500 only becomes cost-neutral against Pro 100 or Pro 200 when API spending reaches $500 a month, and its distinctive features are Ultrafast mode and the largest usage allowance.
plans = {"Plus": 20, "Pro 100": 100, "Pro 200": 200, "Pro 500": 500}
credits = {"Plus": 0, "Pro 100": 100, "Pro 200": 200, "Pro 500": 500}
for api in [0, 50, 100, 200, 400]:
print(api, {p: plans[p] + max(0, api - credits[p]) for p in plans})··········
WHAT OPENAI DOES NOT PUBLISH ABOUT CHATGPT PLUS AND PRO LIMITS.
Several numbers a buyer would want are missing from OpenAI's current pages. The exact usage allowance of each Pro tier is not published: the five-times and 20-times multiples date from before the October 29 doubling, and OpenAI has not stated new ones. Weekly limits are mentioned for both plans without figures. The dollar price of ChatGPT credits for individual plans is not listed. Deep research quotas appear only as "Expanded" for Plus and "Maximum" for Pro, and image generation as "Unlimited and faster" for Pro without a number for Plus.
OpenAI's help-center articles on Pro tiers and on the shared ChatGPT and Codex limit could not be retrieved on November 8, so any tier-specific guidance they contain is not reflected here beyond what the pricing documentation and plan pages repeat. The chatgpt.com plan pages also differ in wording from the Codex and ChatGPT Work documentation: the former describe Pro as "Your choice of 3 usage tiers", while the latter names the $100, $200 and $500 prices and the Pro 500 Ultrafast entitlement.
··········
CHATGPT PLUS OR CHATGPT PRO: WHICH PLAN FITS WHICH SUBSCRIBER.
The published facts support a choice by workload. Plus covers the large majority of individual use: daily chat, writing, research, image creation, projects and a few focused Codex sessions a week, with every current GPT-6 model available. Pro 100 is the logical step up for anyone who keeps hitting the Plus five-hour ranges or needs inputs above about 320 pages, and it is the cheapest choice outright for developers whose API bill is $80 a month or more. Pro 200 and Pro 500 are usage purchases for heavy daily Codex and ChatGPT Work users, with Pro 500 the only individual plan that offers Ultrafast mode.
........
Subscriber or workload | Recommended plan | Reason from the published facts |
Everyday chat, writing and research | Plus | All current GPT-6 models for $20 a month |
Occasional coding in Codex | Plus | 15 to 160 GPT-6.1 Sol local messages per 5 hours |
Daily Codex or ChatGPT Work sessions | Pro 100 | No five-hour limit; usage doubled on Oct 29 |
Documents above about 320 pages | Pro 100 | Reasoning input up to about 680 pages |
Developer with $80+ monthly API spend | Pro 100 | $100 API credits offset the higher price |
Heaviest agentic coding all day | Pro 200 | Larger allowance than Pro 100 |
Lowest latency on Astra and 6.1 Sol | Pro 500 | Only individual plan with Ultrafast |
Always-on agent by text message | Pro (any tier) | Dots and US texting beta are Pro-only |
Team admin, SSO or shared billing | Neither | Business features are not in Plus or Pro |
........
For a subscriber who is unsure, the cheapest test is to stay on Plus, buy credits when a limit is reached, and track how often that happens. If purchased credits approach $80 a month, or the work regularly needs more than about 320 pages of input, Pro 100 is the better plan; if the API is already part of the monthly budget, Pro 100 pays for itself from the first month the credits are fully used.
··········
FOLLOW US FOR MORE.
·····
DATA STUDIOS
·····
datastudios.org
