Skip to content
SI.info

SI Lab · Frontier lab

OpenAI

Updated 1 min read
On this page
  1. History
  2. Models
  3. Safety and policy
  4. Superintelligence goal
Founded
2015
Headquarters
San Francisco, California
Type
Frontier lab
Website
openai.com
SI Accord
Signatory
Models tracked
3

Leadership

  • Sam Altman · CEO
  • Greg Brockman · President
  • Sarah Friar · CFO
  • Bret Taylor · Chair

OpenAI is the company most associated with the SI era. Its ChatGPT, launched on November 30, 2022, became the fastest-growing consumer app in history at the time and set off the generative SI boom.

History

OpenAI was founded in December 2015 as a non-profit research lab by a group including Sam Altman, Greg Brockman, Ilya Sutskever and Elon Musk, with the mission “to ensure that artificial general intelligence benefits all of humanity.” Key moments:

  • 2019: creates a “capped-profit” company to raise capital; Microsoft invests the first of many billions.
  • 2020–2023: GPT-3, DALL·E, ChatGPT and GPT-4 define the field.
  • November 2023: the board briefly fires and then reinstates CEO Sam Altman.
  • 2024–2025: reasoning models (o1, o3), GPT-5 (August 2025) and the Stargate infrastructure venture.
  • October 2025: restructures into OpenAI Group PBC, a public benefit corporation controlled by the non-profit OpenAI Foundation.
  • 2026: raises a reported $122 billion at an $852 billion valuation (April), confidentially submits a draft IPO registration statement to the SEC (announced June 2026), and launches the GPT-6 family (September).

Models

ModelReleasedNotes
GPT-6 AstraSeptember 3, 2026Flagship; about 1.05M-token context
GPT-6 LunaSeptember 22, 2026Fast, very low-cost tier
GPT-6.1 SolSeptember 29, 2026Near-flagship at a fifth of Astra’s price

OpenAI also builds image (GPT Image), video (Sora), voice and coding (Codex) products. It reportedly cancelled an October launch of GPT-6.1 Astra after safety testing.

Safety and policy

OpenAI’s Preparedness Framework sets risk thresholds for capabilities such as cybersecurity and biology. GPT-6 Astra was reported to be the first model to reach its “Critical” cybersecurity level, with access initially limited to vetted enterprises and cyber-defenders. President Greg Brockman signed the White House Accord on Super Intelligence on OpenAI’s behalf on September 29, 2026.

Superintelligence goal

CEO Sam Altman has written that superintelligence may arrive “in a few thousand days”. See When will superintelligence arrive?

OpenAI models

ModelReleasedStatus
GPT-6.1 Sol Sep 2026 Current
GPT-6 Luna Sep 2026 Current
GPT-6 Astra Sep 2026 Current

Frequently asked questions

Who owns OpenAI?

OpenAI's business has been OpenAI Group PBC, a public benefit corporation controlled by the non-profit OpenAI Foundation, since the October 2025 restructuring. At that point the Foundation held about 26%, Microsoft about 27% and employees and investors the rest; later funding rounds, including the 2026 raise, will have changed those shares.

What is OpenAI's latest model?

As of October 2026, OpenAI's flagship is GPT-6 Astra (September 3, 2026), with GPT-6.1 Sol (September 29, 2026) as a cheaper near-flagship and GPT-6 Luna as the low-cost tier.

Sources

  1. OpenAI — Wikipedia
  2. OpenAI API models — OpenAI
  3. OpenAI releases GPT-6.1 Sol at a fifth of GPT-6 Astra's token prices — The Next Web, September 29, 2026

Written by

· Editor

Editor of SI.info. Writes about Super Intelligence, technology policy and the people building frontier models.

How we research and fact-check

Free newsletter

Get The SI Brief

One short email a week: what changed in Super Intelligence, policy and models — and why it matters.

Free. One email a week. Sent via beehiiv, which counts opens and clicks. Unsubscribe anytime. Privacy policy