Skip to content
Search Sign in List your company
Software Developers Artificial Intelligence

OpenAI

The company behind ChatGPT and the OpenAI API. OpenAI builds general-purpose AI models, and the platform that lets other software build on them.

No reviews yet San Francisco, United States Enterprise Founded 2015
Contact
Save
Website
Page last updated
22 August 2026
What these mean

Report a problem with this company

About OpenAI

OpenAI started in December 2015 with a mission it still states plainly: make sure artificial general intelligence benefits everyone.

Most people meet OpenAI through ChatGPT. It runs in a browser, on the desktop and on a phone, and there is a plan for most situations, from a free account through to rollouts across a whole company or university.

Developers meet it through the OpenAI API. That is where the same models are open to build on, whether the job is text and reasoning, generating images, turning speech into text, or holding a live voice conversation. Codex sits alongside that for writing and reviewing code.

The company reorganised in October 2025. The nonprofit is now the OpenAI Foundation, and it controls OpenAI Group PBC, a public benefit corporation that has to answer for the mission as well as the business. The Foundation holds about 26 per cent of the Group. The older capped-profit arrangement, which put a ceiling on what investors could make, no longer applies.

OpenAI is based in San Francisco.

Brands (2)

ChatGPT 3 products OpenAI API 3 products

Products (6)

ChatGPT Business

A shared workspace a team can set up themselves, with the admin and workspace controls that come with working together. This plan used to be called ChatGPT Team.

ChatGPT Edu

The plan universities use to bring ChatGPT to staff and students across a campus.

ChatGPT Enterprise

Built for rolling ChatGPT out across a large organisation, with the admin controls, security and data-privacy commitments that tends to require.

Realtime API

Speech in, speech out, fast enough to hold a conversation. It is what voice agents are built on, and there are variants for live translation and for straight transcription.

GPT-6 Astra

On 1 August 2026 OpenAI published a post titled "Ten advances in mathematics and theoretical computer science". Inside that post the company said the work came from an internal version of Astra, described as "our next major model". That is how the name became public. There was no launch event and no product page. What OpenAI actually stated is narrow. The post covers ten results on open problems in mathematics and theoretical computer science, and each argument was formalised in a Lean certificate so anyone with the Lean compiler can check it without trusting OpenAI. OpenAI also said the tokens used to reach those results would have cost roughly $2,000 at GPT-5.6 Sol API rates. That figure is a research statistic about the work, not a price for the model. Almost everything else is open. OpenAI has published no release date, no pricing, no context window, no model card, no tool support list and no word on which ChatGPT plans would get it. The company has also not said whether the model ships under the Astra name, as GPT-6, or as a point release inside the GPT-5 line. Plenty of reporting describes how Astra is supposed to work internally. None of that came from OpenAI, so none of it is repeated here. The current generation in this line is GPT-5.6, which OpenAI made generally available on 9 July 2026 across ChatGPT, Codex and the API. It ships as three models: Sol as the flagship, Terra as the everyday production option, and Luna as the fastest and cheapest.

GPT-Live-1 in the API Announced

OpenAI introduced GPT-Live on 8 July 2026. Two models arrived, GPT-Live-1 and GPT-Live-1 mini, and they went to ChatGPT users on web, iOS and Android, paid plans first. OpenAI describes them as voice models that can listen and speak at the same time. Developers did not get them. OpenAI's developer account said plainly: "We're bringing GPT-Live-1 and GPT-Live-1 mini to the API soon", and pointed at a sign-up form. OpenAI hosts that form itself at openai.com/form/gpt-live-1-in-the-api. An OpenAI engineering post on 3 August 2026 also refers to the same architecture underpinning an upcoming GPT-Live API. That is the whole of the announcement. "Soon" is the only timing OpenAI has given. There is no date, no price, no rate limit and no endpoint documentation. What developers can use today on this line is the existing Realtime API. As of early August 2026 its published model list covers gpt-realtime-2.1, gpt-realtime-2.1-mini, gpt-realtime-translate and gpt-realtime-whisper. No GPT-Live entry appears in the public API catalogue.

Client reviews

No reviews yet

Be the first to review OpenAI.

Sign in to review

Frequently asked questions

What is OpenAI?
OpenAI is an AI company, started in December 2015. Its mission is to make sure artificial general intelligence ends up benefiting everyone. Most people know it for ChatGPT; developers know it for the OpenAI API.
Is OpenAI a non-profit or a for-profit company?
Both, and the relationship between them is the point. The non-profit OpenAI Foundation controls OpenAI Group PBC, which is a public benefit corporation, so it is legally on the hook for the mission and not only for the returns. The Foundation holds about 26 per cent of the Group. This came out of a reorganisation finished in October 2025, and it replaced the older capped-profit setup.
What is the difference between ChatGPT and the OpenAI API?
ChatGPT is the finished thing you use. The API is what you build with. If you want to ask a question, you open ChatGPT. If you want your own software to ask the question, you call the API.
Which ChatGPT plans are available?
Free, Go, Plus and Pro are the individual plans. Business and Enterprise are for organisations, and Edu is for universities. If you are looking for the plan called Team, that is now Business.
Can I use OpenAI models in my own product?
Yes, that is what the API is for. It covers text and reasoning, images, speech to text and live voice, and you pay for what you use rather than per seat. The current model list and pricing sit in the platform documentation, because both change often enough that any figure written here would go out of date.
Where is OpenAI based, and when did it start?
San Francisco, California, and December 2015.

Independent coverage of OpenAI from the Brandligo blog.

Cookies on Brandligo

We need a couple of cookies to sign you in and keep forms safe, and those are always on. Beyond that we would like to use Google Analytics to see which pages get read, and Google AdSense to show advertising. Neither is loaded unless you say yes, and saying no costs you nothing on this site. What each one does.