ChatGPT Launch Anniversary: Google’s Code Red and the AI Race

ChatGPT turns four Race

ChatGPT’s fourth anniversary falls on 30 November 2026. The landscape it created looks nothing like the one it entered: OpenAI reports more than a billion weekly users, Google more than a billion monthly users for the Gemini app, and both have rebuilt their products around AI assistants that can search, reason and act.

GPT-6 and ‘Intelligent UI’ reach ChatGPT OpenAI

OpenAI brought its GPT-6 models to ChatGPT with a feature it called Intelligent UI: answers that combine text, visuals and interactive elements, from side-by-side comparisons and diagrams to calculators and simple games built inside the conversation. GPT-6 can also start answering while it is still reasoning, adding to the response as it goes.

It is the clearest sign yet of where the contest has moved: from who has a chatbot to who can turn a question into a finished, usable result.

ChatGPT reaches 1.2 billion weekly users OpenAI

Addressing developers, OpenAI described ChatGPT as a shared surface where people and agents work together and said developers could reach its ‘1.2B weekly users’. The figure was a third higher than seven months earlier. The original research preview had needed five days to reach one million.

Gemini app passes one billion monthly users Google

Google said the Gemini app had crossed one billion monthly active users, making it the fastest-growing product in the company’s history. In July, reporting second-quarter results, Alphabet had cited 950 million, alongside ‘Google Search & other’ revenue of $63.27 billion, up 17% from a year earlier.

Three and a half years after the Code Red, Google could point to both a chatbot audience comparable to ChatGPT’s and a search business that was still growing.

Google I/O 2026: Gemini 3.5 and a new search box Google

At I/O, Google launched Gemini 3.5 Flash, which it said beat Gemini 3.1 Pro on coding, agentic and multimodal benchmarks at Flash prices, and introduced Gemini Omni models that generate video. It also showed a redesigned AI search box that accepts images, files, video and Chrome tabs. Gemini 3.5 Pro was promised for the following month but had not reached general availability by mid-August.

GPT-5.5 released OpenAI

OpenAI released GPT-5.5, with Thinking and Pro versions for paying subscribers and a faster Instant model as ChatGPT’s default. The cadence of releases, roughly every few months on both sides, had become part of the competition.

900 million weekly users and a record funding round OpenAI

OpenAI reported 900 million weekly users and 50 million paying subscribers as it announced a funding round of $110 billion, with Amazon, Nvidia and SoftBank as the largest investors. The subscription tier launched in February 2023 had become a mass-market product.

Ads arrive in ChatGPT OpenAI

OpenAI began showing labelled ‘sponsored’ products below some ChatGPT answers for logged-in US adults on the Free and Go plans. It said ads would not influence answers, conversations would not be sold to advertisers, and ads would not appear for under-18s or next to sensitive topics such as health and politics.

The move was significant for the Google story: ChatGPT was now competing not only for queries but for the advertising budgets that pay for Search.

Apple picks Gemini to power the new Siri Google

In a joint statement, Apple and Google said the next generation of Apple Foundation Models would be based on Gemini and Google’s cloud technology. Apple said it had determined that Google’s AI provided ‘the most capable foundation’. ChatGPT had been integrated into Apple Intelligence since late 2024; the new deal put Gemini at the centre of Siri.

OpenAI declares its own ‘code red’ Race

Three years after Google’s alarm, the roles reversed. In an internal memo first reported by The Information, Sam Altman told OpenAI staff that ‘we are at a critical time for ChatGPT’ and called a code red to improve the product, delaying other work including advertising. The trigger was Google’s Gemini 3, released two weeks earlier, which led several benchmarks.

OpenAI released GPT-5.2 on 11 December 2025. ChatGPT still had far more weekly users than Gemini had monthly users, but the gap was closing.

Gemini 3 launches Google

Google released Gemini 3 directly into the Gemini app and, for the first time on launch day, into AI Mode in Search. The model’s reception, strong across reasoning, coding and multimodal tests, is what prompted OpenAI’s internal alarm two weeks later.

The race comes to India India

India became a front line. Reliance Jio and Google announced a free 18-month Google AI Pro offer for eligible Jio users, and days later OpenAI made its ChatGPT Go plan free for a year to users in India who signed up during a promotion. The aim on both sides was the same: win the habit of a very large, young, mobile-first audience early.

800 million weekly users and apps inside ChatGPT OpenAI

OpenAI reported 800 million weekly users and let developers build apps that run inside ChatGPT. Two weeks later it launched ChatGPT Atlas, a web browser built around the assistant, a direct move onto ground long held by Google Chrome.

Judge cites AI chatbots in Google remedies ruling Antitrust

A year after finding that Google had illegally maintained a search monopoly, Judge Amit Mehta declined the US government’s request to force a sale of Chrome. He pointed in part to the rise of generative AI as a new competitive force in search, while barring Google from exclusive distribution contracts and requiring it to share some search data with qualified rivals.

GPT-5 becomes ChatGPT’s default OpenAI

OpenAI launched GPT-5, combining fast responses and deeper reasoning behind one model choice and making it available to free users. It reported about 700 million weekly users at the time.

AI Mode comes to US Search Google

Google opened AI Mode, a chatbot-style search tab powered by Gemini, to all users in the US and reported 400 million monthly users for the Gemini app. Search itself was being rebuilt in the image of the chatbot that had threatened it.

ChatGPT search and Gemini 2.0 Race

OpenAI added real-time web search with linked sources to ChatGPT, rolling it out to all users by early 2025. Google answered in December with Gemini 2.0, pitched as a model for agents that can take actions. The products were converging: chatbots were adding search, and search was adding chat.

Google ruled a search monopolist Antitrust

In a landmark antitrust ruling, Judge Mehta found that Google had illegally maintained a monopoly in general search and search text advertising, largely through payments to be the default search engine on browsers and phones. AI chatbots were still a small share of searches, but they would become central to the remedies argument.

GPT-4o and AI Overviews, a day apart Race

OpenAI showed GPT-4o, a faster ‘omni’ model with natural real-time voice, the day before Google I/O. At I/O, Google rolled out AI Overviews, AI-written summaries above search results, across the US. Within days, screenshots of odd answers, including a suggestion to add glue to pizza sauce, spread online, and Google said it had made more than a dozen technical fixes.

Bard becomes Gemini Google

Google retired the Bard name and gave its assistant the same name as its model family. It launched a Gemini app and a paid tier, Gemini Advanced, running Gemini Ultra 1.0. Weeks later, on 22 February, it paused Gemini’s image generation of people after historically inaccurate images drew criticism, a reminder that the rush to ship carried risks.

Google announces Gemini Google

Google DeepMind introduced Gemini, a family of models designed from the start to handle text, code, images, audio and video, in three sizes for data centres down to phones. A promotional video drew criticism when Google acknowledged it had been edited and did not show real-time interaction. The model nonetheless became the foundation of Google’s response.

OpenAI’s first DevDay, and a boardroom crisis OpenAI

A year after launch, OpenAI reported 100 million weekly users and announced custom GPTs. Eleven days later its board removed Sam Altman as chief executive, saying he had not been consistently candid. After most staff threatened to leave and Microsoft offered to hire them, he returned within five days under a new board.

Google I/O: PaLM 2 and AI in Search Google

Google’s 2023 developer conference was devoted to AI. It announced PaLM 2, moved Bard onto it, dropped the waitlist and previewed the Search Generative Experience, the experiment that became AI Overviews. In April it had merged its two AI labs, Google Brain and DeepMind, into Google DeepMind.

GPT-4 launches; Bard opens a week later Race

OpenAI announced GPT-4, reporting large gains on professional and academic exams, with image input initially limited to a partner. A week later Google opened Bard to the public in the US and UK through a waitlist, describing it as an experiment that could make mistakes.

Bard announced, Bing strikes, a $100 billion stumble Race

Google announced Bard, an experimental chatbot built on a lighter version of LaMDA. A day later, Microsoft unveiled a new Bing with a chat assistant built on OpenAI models. Then Bard’s own promotional example went wrong: it said the James Webb Space Telescope took the first picture of an exoplanet, which a ground-based telescope had done in 2004. Alphabet lost roughly $100 billion in market value on 8 February.

ChatGPT Plus: the experiment starts charging OpenAI

OpenAI introduced ChatGPT Plus, its first paid consumer tier. By then analysts at UBS estimated that ChatGPT had reached about 100 million monthly users in January, roughly two months after launch, faster than any consumer app before it.

Microsoft doubles down; Google calls its founders back Race

Microsoft announced a new multi-year, multibillion-dollar investment in OpenAI, reported at about $10 billion. At Google, The New York Times reported that founders Larry Page and Sergey Brin, long removed from day-to-day work, had been brought in to review AI plans. The same month Alphabet announced 12,000 job cuts while saying it would sharpen its focus on AI.

Google’s ‘Code Red’ Google

The New York Times reported that Google management had declared a ‘code red’ in response to ChatGPT. Sundar Pichai took part in meetings on AI strategy and redirected work across several groups. The worry was not only that a chatbot could answer questions, but that direct answers could reduce the clicks on search ads that funded the company.

Google had the research and a conversational model of its own. Its hesitation, over accuracy, bias and reputation, now looked like a competitive risk.

One million users in five days OpenAI

Demand far outran OpenAI’s expectations. Users shared screenshots of essays, code, poems and explanations, and the service frequently showed an ‘at capacity’ message. For comparison, Instagram had taken about two and a half months to reach one million users.

OpenAI launches ChatGPT OpenAI

OpenAI released ChatGPT, a model it said could ‘answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests’. The underlying technology was not new, but the interface was: anyone could hold a conversation with a large language model in a browser, free.

OpenAI listed limitations from the start, including that it ‘sometimes writes plausible-sounding but incorrect or nonsensical answers’.

Google’s chatbot that stayed in the lab Research

Google showed LaMDA, a model for open-ended dialogue, at I/O 2021 and later offered limited demos through its AI Test Kitchen app. Public attention came mostly from a Google engineer’s claim in June 2022 that LaMDA was sentient, which Google rejected. The company kept the chatbot away from a general release.

OpenAI builds the GPT series Research

OpenAI applied the Transformer to large-scale ‘generative pre-training’. GPT-2 was initially withheld in full over misuse concerns. GPT-3, with 175 billion parameters, was offered to developers through an API in 2020 and later fine-tuned into the instruction-following models behind ChatGPT.

Google invents the Transformer Research

Google researchers published the paper that introduced the Transformer, an architecture that uses attention to model relationships between words across a whole sequence at once. It became the foundation for nearly every large language model, including OpenAI’s GPT series and Google’s own Gemini.