GPT4: The Amazing New AI System from OpenAI that Beats GPT3.5

OpenAI, the research organization dedicated to creating artificial general intelligence (AGI) that benefits all of humanity, has announced its latest breakthrough: GPT4, a large multimodal model that can accept image and text inputs and emit text outputs. GPT4 is the successor of GPT3.5, which was released a year ago as a test run of OpenAI’s new deep learning stack and supercomputer.

What is the difference between GPT4 and GPT3.5?

GPT4 is more reliable, creative, and able to handle much more nuanced instructions than GPT3.5 and other AI chatbots. It also exhibits human-level performance on various professional and academic benchmarks, such as passing a simulated bar exam with a score around the top 10% of test takers; in contrast, GPT3.5’s score was around the bottom 10%.


GPT4 is also more aligned and safer than GPT3.5, thanks to six months of iterative alignment using lessons from OpenAI’s adversarial testing program and ChatGPT, its conversational AI product. According to OpenAI, GPT4 is 82% less likely to respond to requests for disallowed content and 40% more likely to produce factual responses than GPT3.5 on their internal evaluations.

How can I use GPT4?

OpenAI is releasing GPT4’s text input capability via ChatGPT and the API (with a waitlist). ChatGPT is a platform where users can interact with GPT4 on various creative and technical writing tasks, such as composing songs, writing screenplays, or learning a user’s writing style. The API is a service that allows developers to access GPT4’s capabilities via a simple interface.

To prepare the image input capability for wider availability, OpenAI is collaborating closely with a single partner to start. This means that GPT4 can also process images as inputs and generate relevant text outputs, such as captions, descriptions, or stories1.

OpenAI is also open-sourcing OpenAI Evals, its framework for automated evaluation of AI model performance, to allow anyone to report shortcomings in its models and help guide further improvements.

How much does it cost to use GPT4?

OpenAI has announced a new pricing model for GPT4, in which it has reduced the price of the prompt tokens (the words or characters that are given as inputs to GPT4). The price of the sampled tokens (the words or characters that are generated by GPT4 as outputs) remains the same as before.

For its models with 8k context lengths (e.g. gpt-4 and gpt-4-0314), the price is:

  • $0.03/1k prompt tokens
  • $0.06/1k sampled tokens

For its models with 32k context lengths (e.g. gpt-4-32k and gpt-4-32k-0314), the price is:

  • $0.06/1k prompt tokens
  • $0.12/1k sampled tokens

OpenAI says that its pricing model is designed to cover the costs of running and maintaining GPT4, as well as to support its research and development efforts. It also offers discounts for academic or non-profit users.


GPT4 is OpenAI’s most advanced system, producing safer and more useful responses than ever before. It can solve difficult problems with greater accuracy, thanks to its broader general knowledge and problem solving abilities. It can also generate, edit, and iterate with users on creative and technical writing tasks, thanks to its enhanced creativity and collaboration skills.

GPT4 is available via ChatGPT and the API (with a waitlist) for text input tasks, and via a single partner for image input tasks. OpenAI has also open-sourced OpenAI Evals for evaluating its model performance.

GPT4’s pricing model depends on the number of prompt tokens and sampled tokens used by the users, with a discount for academic or non-profit users.

If you are interested in using GPT4 or learning more about it, you can visit OpenAI’s website or join their waitlist.

Leave a Comment

Your email address will not be published. Required fields are marked *