Summary of AI models that can be used with ChatGPT by plan, OpenAI announces details of usage restrictions for o3, o4-mini, and o4-mini-high in ChatGPT

by Focal Foto
OpenAI has announced the usage restrictions for ChatGPT for o3 , o4-mini , and o4-mini-high, which were announced in April 2025. Depending on the plan you subscribe to with your ChatGPT account, you may be limited to 50 messages per week. We have summarized the features and restrictions of the AI models that can be used with ChatGPT at the time of writing for each plan.
OpenAI o3, o4-mini, and o3-mini Usage Limits on ChatGPT and the API | OpenAI Help Center
https://help.openai.com/en/articles/9824962-openai-o3-o4-mini-and-o3-mini-usage-limits-on-chatgpt-and-the-api#h_0151d07654
◆About ChatGPT usage plans
ChatGPT has four plans: a free version that is free to use, and monthly paid subscription plans: Plus, Pro, Team, and Enterprise. Plus costs $20 (about 2,900 yen) per month, Pro costs $200 (about 29,000 yen) per month, Team for business use costs $25 (about 3,600 yen) per person per month if paid annually, $30 (about 4,300 yen) per person per month if paid monthly, and Enterprise is negotiable.
◆ Model available for free users
ChatGPT's free plan includes three types of GPT: GPT-4o, GPT-4o-mini, and o4-mini. If you are a free user, there are strong restrictions on input and generation functions such as images, audio, and files other than text, and file uploads are limited to three times every 24 hours, and image generation is limited to three images per 24 hours. The context window is 8K.
Free Plan | Features | limit |
---|---|---|
GPT-4o | Multimodal support (text, image, audio) High speed and high accuracy | There are certain restrictions (limit values are not disclosed) |
GPT-4o-mini | A lightweight version of GPT-4o. Low-cost, high-speed, and suitable for light tasks. | Unlimited use that switches to 4o-mini when 4o usage limit is reached |
o3-mini | A lightweight version of o3. It responds very quickly and has stable inference. Suitable for general use in a browser. | There are certain restrictions (limit values are not disclosed) |
The free version limits the number of times models can be used, has limitations on bandwidth and availability of responses, and does not offer Deep Research, an intelligence gathering feature for AI agents.
◆Models available for Plus users
The cheapest paid plan, Plus, allows you to use GPT-4o, o3, o4-mini, and o4-mini-high, and you can also select the model. GPT-4o can be used with a loose limit of 80 times per 3 hours. However, it will be limited if abuse or violation of the terms is detected. The context window is 32K.
For Plus users, o3 is limited to 50 messages per week, o4-mini to 150 messages per day, and o4-mini-high to 50 messages per day. In addition, some Plus users can use GPT-4.5 as a research preview, with a usage limit of about 50 messages per week. However, OpenAI has not clearly indicated the usage limit for GPT-4.5, and a warning will be displayed when the number of usages approaches the limit. In addition, 'GPT-4' can be used as a legacy model at the time of writing, but it will be unavailable after April 30, 2025.
Plus Plan | Features | Usage restrictions |
---|---|---|
GPT-4o | Multimodal support (text, image, audio) High speed and high accuracy | 80 messages per 3 hours |
GPT-4o-mini | A lightweight and fast version of GPT-4o. Cost-effective and designed for simple processing and everyday conversation. | Nearly unlimited |
o3 | Large-scale inference model with high accuracy and long text processing capability | Up to 50 messages per week |
o4-mini-high | High-performance version of the o4-mini for coding and visual tasks | Up to 50 messages per day |
GPT-4.5 | Research preview model for enhancing creativity and multilingual support | Available to select users only Approximately 50 messages per week (subject to change) |
o4-mini | High speed and lightweight model, ideal for everyday use | Up to 150 messages per day |
In addition, the Plus plan also offers the video generation AI Sora, but there is a limit to the number of credits that can be used each month, with a maximum resolution of 720p and a length of 5 seconds, up to 1,000 credits available per month.
◆Model that can be used by Pro users
The Pro plan costs about 30,000 yen per month, which is quite expensive for an individual plan, but it is a plan aimed at professionals such as researchers, companies, and AI engineers, and is a full upgrade from the Plus plan. Access to o3, o4-mini, and GPT-4o is ' almost unlimited ,' but it is said that usage may be temporarily restricted based on the terms of use.
Pro Plan | Features | Usage restrictions |
---|---|---|
GPT-4o | Multimodal support (text, image, audio) High speed and high accuracy | Nearly unlimited |
o3 | Large-scale inference model with high accuracy and long text processing capability | Nearly unlimited |
o4-mini | High speed and lightweight model, ideal for everyday use | Nearly unlimited |
o4-mini-high | High-performance version of the o4-mini for coding and visual tasks | Nearly unlimited |
GPT-4.5 | Research preview model for enhancing creativity and multilingual support | Nearly unlimited |
GPT-4o-mini | A lightweight and fast version of GPT-4o. Cost-effective and designed for simple processing and everyday conversation. | Nearly unlimited |
Pro users' requests are given priority, so there are fewer delays even during busy times, and they have the advantage of early access to new features and next-generation models, as well as priority support services. Sora's usage limits are also wider than Plus, and you can generate up to 10,000 credits per month for videos up to 20 seconds in 1080p.
The Team plan, which is intended for corporate use, is the same as the Plus plan, and the Enterprise plan is the same as the Pro plan. However, the company's prompts and data are not used to train the model, and the user has control over them. In addition, data in transit or at rest is encrypted, and a management console and data analysis are available for team use. A 32K token context window is also provided, allowing for longer inputs and processing of larger files. The Enterprise plan requires contacting OpenAI, and pricing is negotiable.
Related Posts:
in Software, Web Service, , Posted by log1i_yk