OpenAI today announced the availability of GPT-4, its next generation large language model that offers several improvements over the last gen GPT-3.5. For example, the GPT-4 can accept 25,000 words of text, images and offer more accurate responses. OpenAI also claims that GPT-4 surpasses ChatGPT in its advanced reasoning capabilities and it scores higher in approximate percentiles among test-takers.
GPT-4 comes with the following improvements:
- GPT-4 is more creative and collaborative than ever before. It can generate, edit, and iterate with users on creative and technical writing tasks, such as composing songs, writing screenplays, or learning a user’s writing style.
- GPT-4 can accept images as inputs and generate captions, classifications, and analyses.
- GPT-4 is capable of handling over 25,000 words of text, allowing for use cases like long form content creation, extended conversations, and document search and analysis.
We spent 6 months making GPT-4 safer and more aligned. GPT-4 is 82% less likely to respond to requests for disallowed content and 40% more likely to produce factual responses than GPT-3.5 on our internal evaluations.
Like previous generation GPTs, GPT-4 was trained on Microsoft Azure AI supercomputers. Also, OpenAI is using Azure AI-optimized infrastructure to deliver GPT-4 to users around the world.
OpenAI also mentioned that they incorporated more human feedback and they wil lbe updating and improving GPT-4 at a regular cadence as more people use it. OpenAI used GPT-4 to help create training data for model fine-tuning and iterate on classifiers across training, evaluations, and monitoring.
You can now access GPT-4 via ChatGPT Plus and also via API to build applications and services.