Research on bioscience capabilities of AI models and safe use in laboratory settings

Los Alamos National Laboratory Collaboration

Development of AI supercomputers, including the 'Stargate' project planned for 2028

Microsoft Partnership

New series of AI models designed to spend more time thinking before responding

OpenAI o1

Prototype search engine combining traditional search with generative AI capabilities

SearchGPT

Updated version of GPT-4 with enhanced capabilities

GPT-4o

Advanced video generation model described as a 'world simulator'

Sora

Research aimed at enhancing AI reasoning capabilities for autonomous planning and deep research

Project Strawberry (Q*)

Ongoing project focusing on robotic hand manipulation

Dactyl

AI system for generating images from text descriptions

DALL-E

AI system for translating natural language into code

Codex

OpenAI, founded in 2015, is an AI research laboratory focused on ensuring artificial general intelligence (AGI) benefits humanity. It has made breakthroughs in machine learning, particularly in natural language processing, generative AI, and reasoning capabilities. OpenAI developed AI models such as GPT-4, DALL·E, Codex, Sora, and the o1 series. The organization combines cutting-edge research, AI safety and ethics, and strategic partnerships to advance AI capabilities across text, image, audio, and video generation. Recent work emphasizes AI reasoning and the safe use of multimodal AI models in scientific settings, exemplified by its collaboration with Los Alamos National Laboratory.

OpenAI Headquarters

New office space in the former Uber Headquarters complex

OpenAI Mission Bay Office

OpenAI's research labs focus on developing advanced AI models and techniques, including natural language processing, computer vision, and reinforcement learning. Shown here is its London office.

AI Research Lab

A series of large language models, including GPT-3 and GPT-4, capable of generating human-like text across a wide range of tasks and domains.

GPT (Generative Pre-trained Transformer)

$1 billion investment in 2019, with ongoing collaboration

Sam Altman

Greg Brockman

Ilya Sutskever

Dario Amodei

Main research facility for OpenAI, located in San Francisco, California

Partnership to develop evaluations for safe use of multimodal AI models in scientific settings

OpenAI

GPT-3 recognized as a major advancement in natural language processing

AI Breakthrough of the Year

Listed among the world's most influential companies in 2023

TIME100 Most Influential Companies

For the development of Mask R-CNN, awarded by the International Conference on Computer Vision in 2017

Best Paper Award

Ongoing development and improvement of the open-source deep learning framework

PyTorch Development

Project focused on scaling human-centered machine translation

No Language Left Behind (NLLB)

Various initiatives aimed at using AI for positive societal impact

AI for Social Good

Development of advanced conversational AI systems

BlenderBot

Application of AI in computational biology, including the prediction of 600 million protein structures in 2022

Protein Structure Prediction

Meta's R&D is conducted through Meta AI Research and Reality Labs, which bring together experts to advance artificial intelligence, virtual and augmented reality, and social connection technologies. Meta focuses on addressing real-world problems with its research and uses AI to improve product experiences.

Meta AI Research – Menlo Park

Research lab focusing on fundamental AI research

Meta AI Research – New York

European research center specializing in computer vision and speech recognition

Meta AI Research – Paris

Meta AI Research – London

Meta AI Research – Montreal

Meta AI Research – Tel Aviv

Meta AI Research – Seattle

Meta AI Research – Pittsburgh

Meta AI Research is Meta's fundamental AI research division, exploring the future of machine learning across areas like computer vision, natural language processing, and reinforcement learning.

Meta AI Research

Reality Labs is Meta's division focused on augmented and virtual reality research, including areas like optics, computer vision, audio, graphics, and human-computer interaction.

Reality Labs

Codec Avatar is a research project from Reality Labs that aims to create highly realistic virtual avatars and worlds for telepresence and virtual reality applications.

Codec Avatar

Primary funding source as part of Meta's R&D investment

Meta Platforms, Inc.

Meta AI Research (Facebook AI Research) Hero Image

Michael Abrash

Thomas Reardon

Brendan Iribe

Conducts fundamental AI research across various fields including machine learning, computer vision, and natural language processing

Meta AI Research (formerly FAIR)

Focuses on virtual reality (VR) and augmented reality (AR) hardware and software development, including Quest headsets and platforms like Horizon Worlds

Part of Reality Labs Research, works on advanced display technologies for VR/AR, awarded prizes at SIGGRAPH 2023

Meta Platforms, Inc.

Overview

Video

Featured Content

Meta AI Research

Research Areas

Publications per Year

Patents per Year

Related Content

Meta’s Sapiens vision models bring 3D analysis of humans to the “wild”

Explore More Labs