ChatGPT Pro
Last reviewed
Sources
22 citations
Review status
Source-backed
Revision
v2 ยท 2,045 words
Improve this article
Add missing citations, update stale details, or suggest a clearer explanation.
Suggest edit
CtrlK
Last reviewed
Sources
22 citations
Review status
Source-backed
Revision
v2 ยท 2,045 words
Add missing citations, update stale details, or suggest a clearer explanation.
ChatGPT Pro is the premium consumer subscription tier of ChatGPT, the conversational artificial intelligence product developed by OpenAI. Launched on December 5, 2024 at $200 per month, it gives subscribers unlimited access to OpenAI's best models plus exclusive, extra-compute reasoning modes (first o1 pro mode, later GPT-5 Pro) and the earliest access to new features such as Operator, Sora, and Deep Research. [1][2] At ten times the price of the $20-per-month ChatGPT Plus tier, it is OpenAI's highest-priced individual plan and is positioned for power users such as researchers, engineers, and other professionals who rely on advanced AI daily. [1][3]
OpenAI introduced ChatGPT Pro as a tier offering "the most access to OpenAI's best models and tools." [1] At launch it bundled unlimited use of the company's leading models with an exclusive feature, o1 pro mode, that allocates extra computing power to difficult reasoning problems. The plan was unveiled by chief executive Sam Altman alongside the general availability of the o1 reasoning model on the opening day of the company's twelve-day series of product announcements. [2][4]
ChatGPT Pro is distinct from the underlying models it provides. It should not be confused with o1, OpenAI's reasoning model, or with o1-pro, the specific model variant exposed through the plan's o1 pro mode. ChatGPT Pro is the subscription; o1-pro is one of the models available within it.
ChatGPT Pro launched on December 5, 2024 at $200 per month as the first announcement of OpenAI's "12 Days of OpenAI" campaign, a series of daily product reveals running through mid-December. [1][2] It was the company's first individual plan priced above the $20-per-month Plus tier and remained its most expensive consumer subscription. In April 2026 OpenAI added a second, lower-priced Pro tier at $100 per month, while keeping the original $200 plan in place (see Pricing structure below). [17]
At launch, ChatGPT Pro provided unlimited access to OpenAI o1, the smaller o1-mini, GPT-4o, and Advanced Voice, the spoken conversation mode of ChatGPT. [1][3] Its signature inclusion was exclusive access to o1 pro mode. [1] OpenAI described the audience as people who "use research-grade intelligence" daily and said the plan would expand over time to include "more powerful, compute-intensive productivity features." [1][5]
The following table contrasts the plan with ChatGPT Plus as the two stood at the December 2024 launch.
| Feature | ChatGPT Plus | ChatGPT Pro |
|---|---|---|
| Price | $20 per month | $200 per month |
| Access to o1, o1-mini, GPT-4o | Yes (rate-limited) | Yes (unlimited, per OpenAI's usage policies) |
| Advanced Voice | Yes (with usage limits) | Yes (with higher limits) |
| o1 pro mode | No | Yes (exclusive) |
| Intended user | General paid users | Power users, researchers, engineers |
OpenAI noted that "unlimited" usage is subject to its policies guarding against abuse, such as account sharing and automated or programmatic resale of access. [1] By 2026 the Pro tier had expanded to bundle unlimited messages and uploads, the highest Deep Research allowance (250 runs per month) and agent-mode limits, expanded Projects and custom GPTs, an expanded Codex coding agent, maximum memory and context, and research previews of new features. [17][18]
o1 pro mode is a version of the o1 model that uses additional compute to "think harder" and produce more reliable answers to the hardest problems. [1] OpenAI stated that, in evaluations by external expert testers, o1 pro mode produced more reliably accurate and comprehensive responses than standard o1, with particular strength in data science, programming, and case law analysis. [1] On internal machine learning benchmarks across mathematics, science, and coding, OpenAI reported that o1 pro mode outperformed both o1 and the o1 preview model, and that the gains were most pronounced on the most challenging problems. [1][3]
Because o1 pro mode can take a noticeably long time to respond, the ChatGPT interface displays a progress bar while it works and can send a notification within the app when an answer is ready, allowing users to switch to other conversations in the meantime. [3]
When OpenAI released GPT-5 on August 7, 2025, it made the new model OpenAI's default across all tiers and gave ChatGPT Pro subscribers a higher-compute variant called GPT-5 Pro. [19][20] OpenAI described it as "a souped-up version called GPT-5 Pro that uses additional computational resources to produce better answers," extending the same extra-inference-compute idea that o1 pro mode had pioneered. [20] GPT-5 Pro became the new exclusive reasoning headliner of the Pro tier, succeeding o1-pro and o3-pro in the model picker.
OpenAI reported that, with GPT-5 Pro's extended reasoning, the model set a state-of-the-art score of 88.4% on GPQA Diamond (a benchmark of expert-level science questions) without tools, and reached 42% on the Humanity's Last Exam benchmark when using tools, the strongest results in the GPT-5 family. [19] As with o1 pro mode, GPT-5 Pro can take longer to answer because it spends more time reasoning before responding. [19][20]
OpenAI continued to attach the highest-compute "Pro" reasoning mode to the ChatGPT Pro tier in later releases. GPT-5.2, released December 11, 2025, shipped in three modes (Instant, Thinking, and Pro), and OpenAI stated that "GPT-5.2 Pro takes more reasoning time and compute than GPT-5.2 Thinking," with particular strength in spreadsheet creation, financial modeling, and multi-step project execution. [21][22]
After launch, ChatGPT Pro became the tier that received the earliest or most generous access to a series of new OpenAI capabilities. In several cases features debuted exclusively for Pro subscribers before reaching cheaper tiers. The table below summarizes the major additions and their verified launch dates.
| Date | Feature | Notes for Pro |
|---|---|---|
| Dec 5, 2024 | o1 pro mode | Exclusive to Pro at launch [1] |
| Dec 9, 2024 | Sora video generation | Pro received 10,000 credits, up to 500 priority videos at 1080p and 20 seconds, plus unlimited lower-priority videos without watermarks [6][7] |
| Jan 23, 2025 | Operator agent | Launched as a research preview exclusive to U.S. Pro users [8][9] |
| Feb 2, 2025 | Deep Research | Released to Pro users first, initially limited to 100 queries per month (later raised to 120) [10][11] |
| Feb 27, 2025 | GPT-4.5 | Research preview rolled out to Pro users first, before Plus and Team [12] |
| Jun 10, 2025 | o3-pro | Released to Pro and Team users, replacing o1-pro in the model picker [13][14] |
| Aug 7, 2025 | GPT-5 Pro | Higher-compute GPT-5 variant exclusive to Pro at launch (later extended to Team, Enterprise, and Edu) [19][20] |
| Dec 11, 2025 | GPT-5.2 Pro | Highest-compute mode of GPT-5.2, available to Pro users [21][22] |
Sora, OpenAI's text-to-video model, launched on December 9, 2024, for both Plus and Pro subscribers, with the Pro tier receiving substantially higher limits and resolution. [6][7] Operator, an agent that performs tasks in a web browser running on OpenAI's servers, debuted on January 23, 2025, as a research preview available only to U.S. Pro subscribers, with launch partners including Uber, OpenTable, StubHub, DoorDash, and Etsy. [8][9] Deep Research, a tool that autonomously browses the web and compiles cited reports, launched on February 2, 2025, for Pro users before being extended to Plus, Team, Edu, and Enterprise users on February 26, 2025. [10][11] The research preview of GPT-4.5 reached Pro users on February 27, 2025, ahead of other tiers. [12] On June 10, 2025, OpenAI released o3-pro, a higher-compute version of its o3 reasoning model, to Pro and Team users, where it replaced o1-pro. [13][14]
ChatGPT Pro sits at the top of OpenAI's individual subscription ladder, above the free tier and the $20-per-month ChatGPT Plus plan, and alongside the business-oriented ChatGPT Team and ChatGPT Enterprise plans. Plus offers paid users higher rate limits than the free tier, while Pro removes most usage caps and adds the exclusive Pro reasoning modes and earliest feature access. [1][3]
On April 9, 2026, OpenAI introduced a second Pro tier priced at $100 per month, positioned chiefly for developers who needed heavy Codex coding capacity, and confirmed that the original $200 plan would remain available. [17] OpenAI said the $200 tier provides roughly 20 times the usage limits of Plus and is meant to support a user's "most demanding workflows continuously, even across parallel projects," while the $100 tier offers about five times the Codex capacity of the Plus plan. [17] OpenAI framed the move as undercutting rivals on coding value, stating that "Codex delivers more coding capacity per dollar across paid tiers." [17]
ChatGPT Pro drew immediate attention for its $200 monthly price, described by TechCrunch as "easily OpenAI's priciest plan yet." [3] At ten times the cost of ChatGPT Plus, it was widely characterized as a tier aimed at a narrow set of professional and research users rather than the general public. [3][4]
Alongside the launch, OpenAI announced a grant program awarding ten free ChatGPT Pro subscriptions to medical researchers at leading U.S. institutions, with stated plans to expand grants to other regions and fields. Early recipients included Dr. Catherine Brownstein of Boston Children's Hospital and Harvard Medical School and Dr. Rhoda Au, a dementia researcher at Boston University. [1]
Commentators were initially mixed on whether the plan justified its price, with some early reviewers concluding that features such as Operator were not yet worth the cost on their own. [9] The subsequent addition of Deep Research, GPT-5 Pro, and other exclusive capabilities led some analysts to argue the plan had become more compelling over time. [10][19]
In early January 2025, Sam Altman publicly stated that OpenAI was losing money on the Pro plan. Writing on X (formerly Twitter), he said: "insane thing: we are currently losing money on openai pro subscriptions! people use it much more than we expected." [15][16] In follow-up remarks he indicated that he had personally chosen the $200 price and had expected the plan to be profitable. [16] The losses were attributed to the heavy computing costs of running OpenAI's most capable models, which were used far more intensively by Pro subscribers than the company had anticipated. [15][16]
The episode was widely reported as an illustration of the high inference costs associated with frontier AI models and the difficulty of pricing unlimited access to compute-intensive products. [16]