Skip to content
ContentLora

    Tip: press / anywhere to search.

    organization

    Google DeepMind

    Also known as DeepMind, Google Brain, DeepMind Technologies

    Google DeepMind is the artificial intelligence lab of Google, formed in April 2023 by merging DeepMind and the Google Brain team.[1] It develops Google's Gemini models, and its AlphaFold work earned a 2024 Nobel Prize in Chemistry.[2][3]

    Editor reviewedUpdated Artificial intelligenceBig TechScience
    Key facts

    Google DeepMind is the artificial intelligence research lab of Google. It was formed in April 2023 by combining DeepMind with the Brain team from Google Research.[1] It develops Google’s Gemini models, and in September 2026 it introduced Gemini 4 Argon.[2] Its rivals in building large AI models include OpenAI and Anthropic.[4]

    Origins

    DeepMind was established in 2010, and Google Brain began in 2011 at X, Google’s moonshot unit.[5] DeepMind was founded by Demis Hassabis, Shane Legg and Mustafa Suleyman.[6] In January 2014 Google agreed to buy the London-based company, at a price reported to be more than $500 million.[7]

    In April 2023 Google merged the two groups into Google DeepMind, with Hassabis as chief executive.[1][8] Google said at the time that he would lead the development of its “most capable and responsible general AI systems.”[8] The lab states its mission as building AI “responsibly to benefit humanity.”[9]

    Research record

    Google DeepMind credits its predecessor teams with the DQN reinforcement learning system, AlphaGo and AlphaFold. It also credits Google Brain with the Transformer architecture that underlies modern language models.[10] Other Google Brain work it lists includes the BERT and PaLM language models, the Imagen text-to-image system and the RT-2 robotics model.[11]

    AlphaFold predicts the 3D structure of proteins from their amino acid sequences. Demis Hassabis and John Jumper shared the 2024 Nobel Prize in Chemistry for the work.[3] By October 2024, more than 2 million researchers in 190 countries had used AlphaFold’s predictions, which are freely available through the AlphaFold Protein Structure Database.[12] Google DeepMind says the AlphaFold 2 paper, published in 2021, is one of the most-cited publications of all time.[13]

    The lab’s science work extends beyond AlphaFold.[14] Its WeatherNext 2 weather model, announced in November 2025, beat the previous WeatherNext model on 99.9% of variables and lead times from 0 to 15 days.[15] Its GNoME project, published in Nature in November 2023, identified 2.2 million crystal structures, 381,000 of which it counted as newly discovered stable materials.[16] The AI for science tracker follows newer results.

    Google DeepMind publishes a Frontier Safety Framework, one of several frontier-safety-frameworks. It updated the framework in September 2025 and again to version 3.1 in April 2026, adding a critical capability level for harmful manipulation and protocols for misalignment risks.[17]

    Gemini models

    On September 30, 2026 Google DeepMind introduced Gemini 4 Argon, a model aimed at software engineering, enterprise knowledge work and cybersecurity defense.[2] The announcement was made by Koray Kavukcuoglu, a senior vice president of Google DeepMind and Google’s chief AI architect.[18] It can produce up to 1 million tokens of output in a single response, up from 64,000 for earlier models, which lets it handle longer tasks.[19] Google first released it to cyber defense partners through its Fairwind Program, with paid API customers and Google AI Ultra subscribers to follow.[20] The introductory API price is $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 afterwards.[21] Google said Argon scored significantly higher than OpenAI’s GPT-6 Astra and Anthropic’s Fable and Opus models across several benchmarks; these are the company’s own figures.[4]

    Adoption

    In July 2026, Alphabet chief executive Sundar Pichai said the Gemini app had 950 million monthly active users, and that its daily active users had tripled over the previous year.[22] He also said Google’s model APIs were processing about 22 billion tokens per minute, up from 16 billion a quarter earlier.[23]

    Robotics

    In July 2026 Google DeepMind announced Gemini Robotics 2. The system gives robots whole-body control, fine dexterity and the ability to work in teams, and it runs locally on the robot, so that humanoid robots can walk, handle objects and coordinate with other robots. Google DeepMind describes it as a way to teach robots to complete a broad range of complex tasks.[24]

    Gemini Robotics 2 builds on the Gemini Robotics family. Google DeepMind demonstrated it controlling full humanoids, including walking and manipulation.[25] The robotics tracker covers the wider field.

    Questions readers ask

    Is Google DeepMind part of Google?

    Yes. Google bought London-based DeepMind in 2014 and in April 2023 merged it with the Brain team from Google Research to form Google DeepMind.[7][1]

    Who runs Google DeepMind?

    Demis Hassabis, a co-founder of DeepMind, is its chief executive.[8][6]

    Why did Google DeepMind researchers win a Nobel Prize?

    Demis Hassabis and John Jumper shared the 2024 Nobel Prize in Chemistry for AlphaFold, an AI system that predicts the 3D structure of proteins from their amino acid sequences.[3]

    What is Gemini 4 Argon?

    It is a Gemini model introduced on September 30, 2026 for software engineering, enterprise knowledge work and cybersecurity defense, with a 1 million token output limit.[2][19]

    Sources

    Each numbered claim is a statement we checked against the sources listed with it. Status shows how well established it is.

    1. [1]

      In April 2023 Google combined the Brain team from Google Research and DeepMind into a single group called Google DeepMind. confirmedas of 2023-04-20

    2. [2]

      On September 30, 2026 Google DeepMind introduced Gemini 4 Argon, a model aimed at software engineering, enterprise knowledge work and cybersecurity defense. confirmedas of 2026-09-30

    3. [3]

      Demis Hassabis and John Jumper of Google DeepMind were awarded the 2024 Nobel Prize in Chemistry for their work on AlphaFold, which predicts the 3D structure of proteins. confirmedas of 2024-10-09

    4. [4]

      Google said Gemini 4 Argon scored significantly higher than OpenAI's GPT-6 Astra and Anthropic's Fable and Opus models across several AI benchmarks. confirmedas of 2026-09-30

    5. [5]

      Google DeepMind combines DeepMind, established in 2010, and Google Brain, which began in 2011 at X, Google's moonshot unit. confirmedas of 2026-10-10

    6. [6]

      DeepMind was founded by Demis Hassabis, Shane Legg and Mustafa Suleyman. confirmedas of 2014-01-26

    7. [7]

      In January 2014 Google agreed to buy London-based DeepMind; the price was reported at more than $500 million. reportedas of 2014-01-26

    8. [8]

      Demis Hassabis, a DeepMind co-founder, is chief executive of Google DeepMind. confirmedas of 2026-10-10

    9. [9]

      Google DeepMind states its mission as building AI responsibly to benefit humanity. confirmedas of 2026-10-10

    10. [10]

      Google DeepMind lists among its predecessors' work the DQN reinforcement learning system, AlphaGo, AlphaFold and, from Google Brain, the Transformer architecture used in modern language models. confirmedas of 2026-10-10

      • About Google DeepMind · Google DeepMind · About page, DeepMind and Google Brain achievements (retrieved 2026-10-10)
    11. [11]

      Google DeepMind credits the Google Brain team with developing the BERT and PaLM language models, the Imagen text-to-image system and robotics systems such as RT-2. confirmedas of 2026-10-10

    12. [12]

      As of October 2024, AlphaFold's freely available predictions had been used by more than 2 million researchers in 190 countries. confirmedas of 2024-10-09

    13. [13]

      Google DeepMind says the AlphaFold 2 paper, published in 2021, is one of the most-cited publications of all time. confirmedas of 2024-10-09

    14. [14]

      AlphaFold is an AI system developed by Google DeepMind that has predicted the structure of millions of proteins since 2021. confirmedas of 2026-03-16

    15. [15]

      Google DeepMind's WeatherNext 2, announced in November 2025, uses a Functional Generative Network and surpasses the previous WeatherNext model on 99.9% of variables and lead times from 0 to 15 days. confirmedas of 2025-11-17

    16. [16]

      Google DeepMind's GNoME work, published in Nature in November 2023, used graph networks to identify 2.2 million crystal structures, of which it counted 381,000 as newly discovered stable materials. confirmedas of 2023-11-29

    17. [17]

      Google DeepMind's Frontier Safety Framework, updated in September 2025 and again to version 3.1 in April 2026, added a critical capability level for harmful manipulation, protocols for misalignment risks such as interference with operator control, and lower "tracked capability levels" to spot risks sooner. confirmedas of 2026-04-17

    18. [18]

      Koray Kavukcuoglu is a senior vice president of Google DeepMind and Google's chief AI architect. confirmedas of 2026-09-30

    19. [19]

      Gemini 4 Argon has a 1 million token output limit, up from 64,000 tokens. confirmedas of 2026-09-30

    20. [20]

      Gemini 4 Argon was first rolled out to cyber defense partners through Google's Fairwind Program, with paid API customers and Google AI Ultra subscribers to follow. confirmedas of 2026-09-30

    21. [21]

      Gemini 4 Argon's introductory API price is $2 per million input tokens and $10 per million output tokens, rising to $4 and $20 after the introductory period; cached inputs are discounted 95 percent. confirmedas of 2026-09-30

    22. [22]

      As of July 2026 the Gemini app had 950 million monthly active users, according to Alphabet chief executive Sundar Pichai. confirmedas of 2026-07-22

    23. [23]

      As of July 2026 Google's model APIs were processing about 22 billion tokens per minute, up from 16 billion a quarter earlier. confirmedas of 2026-07-22

    24. [24]

      In July 2026 Google DeepMind announced Gemini Robotics 2, which provides whole-body control, dexterity and multi-robot collaboration, letting humanoid robots walk, manipulate objects and coordinate with other robots while running locally on devices. confirmedas of 2026-07-30

    25. [25]

      On July 30, 2026 Google DeepMind announced Gemini Robotics 2, a vision-language-action model it says can control full humanoids, including walking and manipulation, as well as bi-arm robots; previous models in the family controlled a humanoid's upper body for tabletop tasks. confirmedas of 2026-07-30

    Revision history (2)
    1. Page created.
    2. Refresh: added science, robotics and safety-framework links to the research-field pages.

    Created Oct 10, 2026. Last reviewed by an editor on Oct 10, 2026. Next scheduled review: Jan 10, 2027.

    Cite this page

    "Google DeepMind." ContentLora, updated Oct 10, 2026. https://contentlora.com/wiki/google-deepmind

    Spotted an error? Suggest a correction or emailcorrections@contentlora.com.

    Keep exploring

    Connections
    OpenAI CompetitorAnthropic CompetitorAlphaFold Developer ofGemini Robotics Developer ofWeatherNext Developer ofGNoME Developer of
    ☰ Contents