The Gemini architecture is trained natively on multiple data types, allowing the models to process and generate text, computer code, images, audio, and video simultaneously. Google distributes the technology in varying capacities, ranging from efficient on-device versions ("Nano") and cost-effective, high-throughput variants ("Flash") to high-compute models designed for complex reasoning ("Pro" and "Ultra"). The 1.5 and 3.0 model generations introduced extended context windows, enabling the analysis of large datasets such as entire codebases, long-form videos, or extensive document archives in a single prompt.
Gemini was first announced on December 6, 2023, and replaced existing Google branding for AI services. In February 2024, the Bard chatbot was renamed Gemini, and the "Duet AI" branding forGoogle Cloud and Workspace was retired in favor of the Gemini identifier. The models integrate into the Google ecosystem through the Gemini mobile app, which functions as an overlay assistant on Android devices, and through the Vertex AI platform for third-party developers.
The release of Gemini has generated technical praise and public controversy. Commentators have highlighted the models' benchmarks in coding and retrieval tasks as competitive withOpenAI'sGPT-4 andGPT-5. However, the product launch faced criticism regarding the reliability of its outputs. In early 2024, Google suspended the model's ability to generate images of people after users reported historical inaccuracies and bias in its depictions of human subjects. Subsequent updates, including the Gemini 1.5 and 3.0 series released throughout 2025, focused on reducinghallucinations, improving latency, and enhancing agentic capabilities for autonomous research and software development.
In November 2022,OpenAI launchedChatGPT, a chatbot based on theGPT-3 family oflarge language models (LLMs).[10][11] ChatGPT gained worldwide attention, becoming aviral Internet sensation.[12] Alarmed by ChatGPT's potential threat toGoogle Search,Google executives issued a "code red" alert, reassigning several teams to assist in the company'sartificial intelligence (AI) efforts.[13]Sundar Pichai, the CEO of Google and parent companyAlphabet, was widely reported to have issued the alert, but Pichai later denied this toThe New York Times.[14] In a rare move, Google co-foundersLarry Page andSergey Brin, who had stepped down from their roles as co-CEOs of Alphabet in 2019, attended emergency meetings with company executives to discuss Google's response to ChatGPT.[15] Brin requested access to Google's code in February 2023, for the first time in years.[16]
Google had unveiledLaMDA, a prototype LLM, earlier in 2021,[17][18] but it was not released to the public.[19] When asked by employees at an all-hands meeting whether LaMDA was a missed opportunity for Google to compete with ChatGPT, Pichai andGoogle AI chiefJeff Dean said that while the company's chatbot had similar capabilities to ChatGPT, there were risks to introducing an LLM that might spread false information, so they decided to wait.[20][21]
In January 2023,Google Brain's sister companyDeepMind CEODemis Hassabis hinted at plans for a ChatGPT rival,[22] and Google employees were instructed to accelerate progress on a ChatGPT competitor, intensively testing "Apprentice Bard" and other chatbots.[23][24] Pichai assured investors during Google's quarterly earningsinvestor call in February that the company had plans to expand LaMDA's availability and applications.[25]
On February 6, 2023, Google announced Bard, agenerative artificial intelligence chatbot powered by LaMDA.[26][27][28] Bard was first rolled out to a select group of 10,000 "trusted testers",[29] before a wide release scheduled at the end of the month.[26][27][28] The project was overseen by product lead Jack Krawczyk, who described the product as a "collaborative AI service" rather than asearch engine,[30][31] while Pichai detailed how Bard would be integrated into Google Search.[26][27][28]Reuters calculated that adding ChatGPT-like features to Google Search could cost the company $6 billion in additional expenses by 2024, while research andconsulting firm SemiAnalysis calculated that it would cost Google $3 billion.[32] The technology was developed under the codename "Atlas",[33] with the name "Bard" in reference tothe Celtic term for a storyteller and chosen to "reflect the creative nature of the algorithm underneath".[34][35]
Multiple media outlets and financial analysts described Google as "rushing" Bard's announcement to preempt rivalMicrosoft's planned February 7 event unveiling its partnership with OpenAI to integrate ChatGPT into itsBing search engine in the form of Bing AI (later rebranded asMicrosoft Copilot),[36][37] as well as to avoid playing "catch-up" to Microsoft.[38][39][40] Microsoft CEOSatya Nadella toldThe Verge: "I want people to know that we made them dance."[41] Tom Warren ofThe Verge andDavey Alba ofBloomberg News noted that this marked the beginning of another clash between the twoBig Tech companies over "the future of search", aftertheir six-year "truce" expired in 2021;[36][42] Chris Stokel-Walker ofThe Guardian, Sara Morrison ofRecode, and analyst Dan Ives of investment firmWedbush Securities labeled this an AIarms race between the two.[43][44][45]
After an "underwhelming" February 8 livestream in Paris showcasing Bard, Google's stock fell eight percent, equivalent to a $100 billion loss inmarket value, and theYouTube video of the livestream was made private.[38][46][47] Many viewers also pointed out an error during the demo in which Bard gives inaccurate information about theJames Webb Space Telescope in response to a query.[48][49] Google employees criticized Pichai's "rushed" and "botched" announcement of Bard on Memegen, the company's internal forum,[50] while Maggie Harrison ofFuturism called the rollout "chaos". Pichai defended his actions by saying that Google had been "deeply working on AI for a long time", rejecting the notion that Bard's launch was aknee-jerk reaction.[51]
A week after the Paris livestream, Pichai had 80,000 employees dedicate two to four hours todogfood testing Bard,[52] while Google executivePrabhakar Raghavan had employees correct any errors Bard made.[53][29] In the following weeks, Google employees criticized Bard in internal messages, citing safety and ethical concerns and calling on company leaders not to launch the service. Google executives launched the product, overruling a negativerisk assessment report conducted by its AI ethics team.[54] After Pichai suddenlylaid off 12,000 employees later that month due to slowing revenue growth, remaining workers shared memes and snippets of their humorous exchanges with Bard soliciting its "opinion" on the layoffs.[55] Google employees began testing a more sophisticated version of Bard with larger parameters, dubbed "Big Bard", in mid-March.[56]
The logo for Bard (21 March 2023 - 7 February 2024)
Google opened up early access for Bard on March 21, 2023, in a limited capacity, allowing users in the US and the UK to join a waitlist. Unlike Microsoft's approach with Bing Chat, Bard was launched as a standaloneweb application featuring atext box and a disclaimer that the chatbot "may display inaccurate or offensive information that doesn't represent Google's views". Three responses are then provided to each question, with users prompted to submit feedback on the usefulness of each answer. Google vice presidents Sissie Hsiao and Eli Collins framed Bard as a complement to Google Search and stated that the company had not determined how to make the service profitable.[57][58][59] Among those granted early access were those enrolled in Google's "Pixel Superfans"loyalty program,[60] users of itsPixel andNest devices, andGoogle One subscribers.[61]
Bard is trained by third-party contractors hired by Google, includingAppen andAccenture workers, whomBusiness Insider and Bloomberg News reported were placed under extreme pressure, overworked, and underpaid.[62][63] Bard is also trained on data from publicly available sources, which Google disclosed by amending itsprivacy policy.[64] Shortly after Bard's initial launch, Google reorganized the team behindGoogle Assistant, the company'svirtual assistant, to focus on Bard instead.[65] Google researcher Jacob Devlin resigned from the company after claiming that Bard had surreptitiously leveraged data from ChatGPT;[66] Google denied the allegations.[67]
Meanwhile, a senior software engineer at the company published an internal memo warning that Google was falling behind in the AI "arms race", not to OpenAI but to independent researchers inopen-source communities.[68] Pichai revealed on March 31 that the company intended to "upgrade" Bard by basing it onPaLM, a newer and more powerful LLM from Google, rather than LaMDA.[69]
The same day, Krawczyk announced that Google had added "math and logic capabilities" to Bard.[70] Bard gained the ability to assist in coding in April, being compatible with more than 20 programming languages at launch.[71][72] Microsoft also began running advertisements in theaddress bar of a developer build of theEdge browser, urging users to try Bing whenever they visit the Bard web app.[73]9to5Google reported that Google was working to integrate Bard into itsChromeOS operating system and Pixel devices.[74][75]
Bard took center stage during the annualGoogle I/O keynote in May 2023,[76] with Pichai and Hsiao announcing a series of updates to Bard, including the adoption of PaLM 2, integration with otherGoogle products and third-party services, expansion to 180 countries, support for additional languages, and new features.[77] In stark contrast to previous years, the Google Assistant was barely mentioned during the event.[76] The expanded rollout did not include any nations in theEuropean Union (EU), possibly reflecting concerns about compliance with theGeneral Data Protection Regulation.[78] Those withGoogle Workspace accounts also gained access to the service.[79]
In June, Google attempted to launch Bard in the EU but was blocked by the IrishData Protection Commission, who requested a "data protection impact assessment" from the company.[80] In July, Bard was launched in the EU and Brazil, adding support for dozens of new languages and introducing personalization and productivity features.[81][82] An invite-only chatroom ("server") onDiscord was created in July, consisting of users who heavily used Bard. Over the next few months, the chatroom was flooded with comments questioning the usefulness of Bard.[83]
Google released a major update to the chatbot in September, integrating it into many of its products through "extensions", adding a button to attempt tofact-check AI-generated responses through Google Search, and allowing users to share conversation threads.[84] Google also introduced the "Google-Extended"web crawler as part of its search engine'srobots.txt indexing file to allowweb publishers to opt-out of allowing Bard to scan them for training.[85] Online users later discovered that Google Search wasindexing Bard conversation threads on which users had enabled sharing; Google stated that this was an error which was corrected.[86]
In October, during the company's annualMade by Google event, Hsiao unveiled "Assistant with Bard", an upgraded version of the Google Assistant which was integrated with Bard.[87] When theU.S. Copyright Office solicited public comment on potential new regulation on generative AI technologies, Google joined with OpenAI and Microsoft in arguing that the responsibility forgenerating copyrighted material lay with the user, not the developer.[88] Accenture contractors voted to join theAlphabet Workers Union in November, in protest of suboptimal working conditions,[89] while the company filed a lawsuit in theU.S. District Court for the Northern District of California against a group of unidentified scammers who had been advertisingmalware disguised as a downloadable version of Bard.[90][91]
On December 6, 2023, Google announcedGemini, a larger,multimodal LLM.[95][96] A specially tuned version of the mid-tier Gemini Pro was integrated into Bard, while the larger Gemini Ultra was used for "Bard Advanced" in 2024.[97][98]The Wall Street Journal reported that Bard was then averaging around 220 million monthly visitors.[99] Google ended its contract with Appen in January 2024,[100] while image generation was added to Bard the next month, using Google Brain'sImagen 2text-to-image model.[101]
On February 8, 2024, Bard and Duet AI were unified under the Gemini brand, with a mobile app launched onAndroid and the service integrated into the Google app oniOS. On Android, users who downloaded the app saw Gemini replace Assistant as their device's default virtual assistant, though Assistant remained a standalone service.[102][103] Google also launched "Gemini Advanced with Ultra 1.0", available via a "Google One AI Premium" subscription,[104] incorporated Gemini into itsMessages app on Android,[105][106] and announced a partnership withStack Overflow.[107][108]
Gemini again took center stage at the 2024 Google I/O keynote.[109][110] Google announced that Gemini would be integrated into several products, including Android,Chrome,Photos, and Workspace.[109] Gemini Advanced was upgraded to the "Gemini 1.5 Pro" language model, with Google previewing Gemini Live, a voice chat mode, and Gems, the ability to create custom chatbots.[111][112][113]
Originally introduced by Google in August 2024, Gemini Live debuted on thePixel 9 series later that year as the default virtual assistant, replacing the Google Assistant on those devices. It was later rolled out to Samsung Galaxy devices beginning with theGalaxy S25 series in July 2025,[114][115] and subsequently expanded to models like theZ Fold 7 andZ Flip 7, where it became part of theGalaxy AI suite.[116]
In February 2025, Google introduced a feature for Gemini Advanced subscribers that can recall and reference past chats.[117]
In June 2025, Google announced Gemini CLI[118], which is an open-source AI tool for terminals.[119] At the same time, Google unveiled a new logo of Gemini that align its colors with color schemes used by the new gradientGoogle logo, with more rounded sparkles than the previous logo.[120] The new logo is available for Android and iPhone devices.[121]
Gemini received mixed reviews upon its initial release in March 2023.[122] James Vincent ofThe Verge found it faster than ChatGPT and Bing Chat, but noted that the lack of Bing-esque footnotes was "both a blessing and a curse",[59] encouraging Google to be bolder when experimenting with AI.[123] His colleague David Pierce was unimpressed by its uninteresting and sometimes inaccurate responses,[124] adding that despite Google's insistence that Gemini was not a search engine, itsuser interface resembled that of one, which could cause problems for Google.[125] Cade Metz ofThe New York Times described Gemini as "more cautious" than ChatGPT,[126] while Shirin Ghaffary ofVox called it "dry and uncontroversial" due to the reserved nature of its responses.[127]
The Washington Post columnist Geoffrey A. Fowler found Gemini a mixed bag, noting that it acted cautiously but could show Internet-influenced bias.[128] Writing forZDNET, Sabrina Ortiz believed ChatGPT and Bing Chat were "more capable overall" in comparison to Gemini,[129] whileWired journalist Lauren Goode found her conversation with Gemini "the most bizarre" of the three.[130] After the introduction of extensions,The New York Times'Kevin Roose found the update underwhelming and "a bit of a mess",[131] whileBusiness Insider's Lakshmi Varanasi found that Gemini often leaned more into flattery than facts.[132]
In a60 Minutes conversation with Hsiao, Google senior vice presidentJames Manyika, and Pichai,CBS News correspondentScott Pelley found Gemini "unsettling".[133] Associate professor Ethan Mollick of theWharton School of theUniversity of Pennsylvania was underwhelmed by its artistic ineptitude.[134] TheNew York Times conducted a test with ChatGPT and Gemini regarding their ability to handle tasks expected of human assistants, and concluded that ChatGPT's performance was vastly superior to that of Gemini.[135]NewsGuard, a tool that rates the credibility of news articles, found that Gemini was more skilled at debunking known conspiracy theories than ChatGPT.[136] A report published by theAssociated Press cautioned that Gemini and other chatbots were prone to generate "false and misleading information that threaten[ed] to disenfranchise voters".[137]
Gemini's response when asked to "generate a picture of a U.S. senator from the 1800s" in February 2024, as shown byThe Verge[138]
In February 2024, social media users reported that Gemini was generating images that featuredpeople of color and women in historically inaccurate contexts—such asVikings,Nazi soldiers, and theFounding Fathers—and refusingprompts to generate images of white people. These images were derided on social media, including byconservatives who cited them as evidence of Google's "wokeness".[138][139][140] The business magnateElon Musk, whose companyxAI operates the chatbotGrok, was among those who criticized Google, denouncing itssuite of products as biased and racist.[141][142][143] Musk and other users targeted Krawczyk, resurfacing his past comments discussing race,[142][143] leading Krawczyk to withdraw fromX andLinkedIn.[144] The conservative-leaning tabloidNew York Post ran a cover story on the incident in the print edition of its newspaper.[145][146]
In response, Krawczyk said that Google was "working to improve these kinds of depictions immediately", and Google paused Gemini's ability to generate images of people.[147][148][149] Raghavan released a lengthy statement addressing the controversy, explaining that Gemini had "overcompensate[d]" amid its efforts to strive for diversity and acknowledging that the images were "embarrassing and wrong".[150][151][152] In an internal memo to employees, Pichai called the debacle offensive and unacceptable, promising structural and technical changes.[153][154][155] Several employees in Google's trust and safety team were laid off days later.[156][157] Hassabis stated that Gemini's ability to generate images of people would be restored within two weeks;[158][159][160] it was ultimately relaunched in late August, powered by its new Imagen 3 model.[161][162]
The market reacted negatively, with Google's stock falling by 4.4 percent.[163] Pichai faced growing calls to resign, including from technology analystsBen Thompson andOm Malik.[164][165][166]House Republicans led byJim Jordan subpoenaed Google, accusing the company of colluding with theBiden administration to censor speech.[167][168] In light of the fiasco and Google's overall response to OpenAI,Business Insider's Hugh Langley and Lara O'Reilly declared that Google was fast going "from vanguard to dinosaur".[169] Bloomberg columnistParmy Olson suggested that Google's "rushed" rollout of Gemini was the cause of its woes, not "wokeness".[170]
Martin Peers, writing forThe Information, opined that Google needed a leader likeMark Zuckerberg to defuse the situation.[171]Hugging Face scientistSasha Luccioni andSurrey University professorAlan Woodward believed that the incident had "deeply embedded" roots in Gemini's training corpus and algorithms, making it difficult to rectify.[172] Jeremy Kahn ofFortune called for researchers focused on safety and responsibility to work together to develop better guardrails.[173]New York magazine contributor John Herrman wrote: "It's a spectacular unforced error, a slapstick rake-in-the-face moment, and a testament to how panicked Google must be by the rise of OpenAI and the threat of AI to its search business."[174]
During the2024 Summer Olympics in July, Google aired a commercial for Gemini entitled "Dear Sydney" depicting a father asking the chatbot to generate afan letter to the star athleteSydney McLaughlin-Levrone for his young daughter. Similar toApple's "Crush!" commercial for theseventh-generation iPad Pro, the advertisement drew heavy backlash online, with criticism for replacing authentic human expression and creativity with a computer;[175][176][177]The Washington Post columnistAlexandra Petri lambasted the commercial as "missing the point".[178] As a result, Google withdrew the commercial fromNBC's rotation.[179][180]
Google aired twocommercials duringSuper Bowl LIX in February 2025, both promoting Gemini. The first, entitled "50 States, 50 Stories", consisted of a national spot and 50 regional spots showcasing howsmall businesses in eachU.S. state leverage Gemini in Google Workspace.[181][182] Social media users noticed a factual error inWisconsin's spot regardinggouda cheese, prompting Google to edit out the incorrect statistic,[183][184][185] whileThe Verge claimed that Google had "faked" some of Gemini's output in the same commercial by plagiarizing text on the web.[186][187][188]
Garett Sloane ofAd Age commented that these blunders illustrated the risks of advertising AI technology.[189] The other commercial was entitled "Dream Job" and featured a father using Gemini on his Pixel 9 to prepare for a job interview; Google also ran a third commercial entitled "Party Blitz" online, in which a man "attempts to impress his girlfriend's family by using Gemini [on his Pixel 9] to become a football expert".[190][191][192]
In 2022,McLaren Racing announced a multi-year partnership with Google.[193] As part of Google's partnership extension with McLaren in 2024,[194] Gemini was advertised on the McLarenFormula One car,[195] including racing a special livery based on Gemini's color palette for the2025 United States Grand Prix.[196]
In the aftermath of the image generation controversy, some users began accusing Gemini's text responses of being biased toward thepolitical left. In one such example that circulated online, Gemini said that it was "difficult to say definitively" whether Musk or the Nazi dictatorAdolf Hitler had more negatively affected society.[197][159] Other users reported that Gemini tended to promote left-wing politicians and causes such asaffirmative action andabortion rights while refusing to promote right-wing figures,meat consumption, andfossil fuels.[165][198][199]The Wall Street Journal'seditorial board wrote that Gemini's "apparently ingrained woke biases" were "fueling a backlash toward AI on the political right, which is joining the left in calling for moreregulation."[200]
In November 2024, CBS News reported that Gemini had responded to a college student inMichigan asking for help with homework in a threatening manner, calling the student "a burden on society" and saying "Please die. Please." A statement issued by Google said "This response violated our policies and we've taken action to prevent similar outputs from occurring."[205]
.png)
.jpg)
