OpenAI is a private artificial intelligence research laboratory consisting of the for-profit OpenAI LP and its parent company, the nonprofit OpenAI Inc. The company was founded in December 2015 with the mission of building advanced artificial intelligence capabilities, aspiring to the goal of achieving artificial general intelligence (AGI).

OpenAI conducts research on a wide range of topics related to artificial intelligence, including machine learning, computer vision, natural-language processing, and robotics. The company also develops and releases various AI-related software tools and products, such as GPT-4, Codex, and DALL-E 2. In late 2022, they released ChatGPT as a free consumer offering, creating a surge in interest in their technology.

Learn more about OpenAI.

We've had an existing investment and collaborative relationship with OpenAI since 2019. Our partnership has accelerated the development and commercialization of AI products and services, promoting innovation in technology critical to global competitiveness. In January 2023 we announced the next phase of that relationship. It involves a multiyear, multibillion-dollar commitment from Microsoft to continue to invest in, innovate on, and build an advanced AI computing infrastructure used for the research, development, and commercialization of OpenAI’s models, such as ChatGPT. And it continues to enable Microsoft and OpenAI to commercialize the resulting AI technology and innovations. OpenAI can independently commercialize their models both in first-party products and with third-party customers, and Azure remains the exclusive cloud provider for OpenAI workloads.

Microsoft and OpenAI share a commitment to building AI systems and products that are trustworthy and safe. OpenAI’s leading research on AI alignment and Microsoft’s Responsible AI Standard not only establish a leading and advancing framework for the safe deployment of our own AI technologies but will also help guide the industry toward more responsible outcomes.

A generative pre-trained transformer (GPT) is a type of language model developed by OpenAI. These models are pre-trained on a large corpus of text data and can be fine-tuned for a variety of natural-language processing tasks, such as language translation, text summarization, and question answering. GPT-4 is the latest and most advanced model in the GPT series, which has set a new level of performance on many natural-language understanding benchmarks and has also been used in a variety of industry and research applications.

Codex is an AI-powered code search and completion tool that allows developers to search through code repositories and find examples of how to use specific APIs or programming constructs. It uses a GPT-3 based model to generate natural-language descriptions of code examples and can generate code snippets as well.

DALL-E is a generative model developed by OpenAI. DALL-E generates images from text prompts; it can produce images of objects and scenes that do not exist in the real world but are constructed by the model based on the text description provided. The model can also generate variations of real-world images based on the text prompt. DALL-E is trained on a massive dataset of images and captions, allowing it to learn the relationship between language and visual content. 

ChatGPT is a variant of the generative pre-trained transformer model. It's fine-tuned by conversational interactions and specifically designed for conversational AI applications, such as dialogue systems, chatbots, virtual assistants, customer service chatbots, and conversational agents for gaming and entertainment. 

When a user inputs a statement or question, ChatGPT generates a response by analyzing the input and considering the context of the conversation. Using the patterns learned from the training dataset, ChatGPT generates a response, aiming to be coherent with the conversation's context and also humanlike. ChatGPT can also be fine-tuned on specific tasks, such as answering questions or providing customer service.

Turing and Megatron, large AI models, continue to help drive research. These models power features within Text Analytics and Microsoft Translator services. We'll continue to leverage our model investments like Turing, Megatron, and GPT-4 to deliver Applied AI and Cognitive Services based on customer demand. There is a rapid expansion of models, and Microsoft is at the center of it all, enabling customers to benefit from each model's strengths.