AI Blog: ChatGPT Tour Part 2

Welcome back to our AI blog.  Today, we take a look at the plans available for ChatGPT and some of the features only available to paid users.

Last week, we started to take a more in depth look at some of the features of ChatGPT.  This week, we will take a look at some of the plans and features available to paid users.

ChatGPT

As mentioned previously, ChatGPT is OpenAI’s AI chatbot, accessible at https://chatgpt.com/.  There are four [4] different plans available:

1. Free: the free version has more limited access to the various features and models available in ChatGPT and includes a more restrictive message cap (especially when using the more advanced features).  This can be sufficient for casual usage, but if you’re using ChatGPT regularly or trying to solve any sort of advanced problem then the message caps may begin to hinder you
2. Plus: this paid subscription gives access to further features (including additional models and DALL·E image generation) and higher message caps allowing much more flexible usage
3. Team: as well as removing many of the caps in place on the features available in the Plus subscription, this allows collaborative workspaces for creating and sharing GPTs, includes an admin console for workspace management and (as mentioned last time) excludes your data from training future versions of ChatGPT by default.  This plan provides a higher level of security, being SOC 2 compliant (to be explained at a later date)
4. Enterprise: this is for use across larger organisations.  The plan is similar to the Team plan but allows for more customisation and has better support.

With the Plus plan (or higher) you can choose which model you’re using, let’s consider each of these:

• GPT-3.5: this was the first publicly released model of ChatGPT and is no longer available through the menu shown above
• GPT-4: this version was the upgrade to 3.5, using more advanced reasoning, responding faster and supporting both text and image inputs
• GPT-4o: this is a newer version of GPT-4 that runs faster, using fewer resources.  Note that ‘GPT-4o with canvas’ uses this same version but within ChatGPT’s collaborative canvas environment, which we’ll take a look at next time
• o1-preview: this is the newest model and is very advanced, note that the current usage limit for Plus or Team accounts are 50 messages a week.  This model specialises in complex reasoning (such as mathematics, coding and science, which traditional models often struggle with).  The o1-preview is presently considerably slower than other models but is much more precise
• The mini models: both GPT-4o and the o1 model have a ‘mini’ version, which are simply faster, cheaper but less accurate versions of their corresponding model.

More next time on AI.