TheChek News Flow

Tired of trailing OpenAI, CEO Sundar Pichai wants his Gemini chatbot to be used by 500 million people

Google’s chief executive thinks his company has the best artificial-intelligence technology on the market. Now he has to get consumers to care.

CEO Sundar Pichai recently told employees he believes the company’s Gemini AI technology has surpassed the capabilities of competitors and he wants the chatbot built on it to be used by 500 million people by the end of 2025, said people familiar with the matter.

Google hasn’t said how many people currently use Gemini, but market leader ChatGPT has about 300 million weekly users. The Gemini app was the 54th most downloaded free app on iPhones Wednesday. ChatGPT was No. 4.

Google was caught flat-footed by the launch of ChatGPT in late 2022, despite being one of the pioneers of AI and spending years working on similar chatbots. The company has been scrambling to catch up ever since.

Gemini isn’t yet a serious moneymaker, but it is central to Google’s push to stay on the cutting edge of AI, a technology already starting to reshape the company’s core search and ads businesses.

Google’s newest AI technology jumped ahead of ChatGPT maker OpenAI’s technology last month on the Chatbot Arena rankings closely watched by people in Silicon Valley. OpenAI dominated the rankings for much of the past year.

Google launched the Gemini chatbot, formerly Bard, a few months after ChatGPT and has updated it with dozens of new features in an effort to gain ground. OpenAI, meanwhile, began directly targeting Google last summer when it added a web-search function to ChatGPT.

The ChatGPT mobile app has been downloaded about 465 million times on Android and iOS devices, compared with 106 million for Gemini, according to Sensor Tower data.

Most of those downloads came on Android phones. Until recently, Gemini was part of Google’s search app on Apple devices. A separate Gemini app for iOS made its debut in November. Google said it was downloaded more than 10 million times through the end of the year.

Usage of AI features built into other applications, such as Meta Platforms’ Facebook and Instagram, aren’t captured by publicly available measures.

Until recently, the Gemini chatbot team reported to the executive in charge of search and ads, but a recent reorganization moved it to the Google DeepMind division that does advanced research.

Pichai said at the time that the move would “enable fast deployment of our new models in the Gemini app.”

Growth in search advertising has slowed in recent years, and Google has increasingly looked to new business lines including subscriptions. It offers a premium version of Gemini with more features for $20 a month, similar to other chatbots.

But unlike its competitors, Google packages its Gemini subscription with perks including AI features in apps like Gmail and two terabytes of file storage.

Sifei Han, a data scientist living in Philadelphia, upgraded to the Gemini premium plan last year after being offered a two-month free trial. He said he mainly uses it for the extra storage, not the AI features.

“I’m more familiar with ChatGPT’s style,” said Han, who also pays for ChatGPT Plus.

About 60% of Gemini’s paying users kept their subscriptions six months after first signing up, according to an Earnest Analytics analysis of debit- and credit-card data from millions of anonymous consumers. That was better than some competitors such as Character.AI and Perplexity, but lagged behind OpenAI and Anthropic, which makes the chatbot Claude.

Subscriptions to Gemini were a fraction of the $10.7 billion in revenue Google parent Alphabet GOOGL 3.11%increase; green up pointing triangle made from subscriptions, device sales and purchases in its mobile app store in the three months ended September. Google’s core advertising business brought in $65.9 billion during the same period. The company hasn’t introduced ads in Gemini.

Gemini has been a source of controversy in the past. Google apologized to users and rolled back an image-generation feature last year after the chatbot produced images of Black soldiers in Nazi garb and sometimes refused to generate depictions of white people. It restored the feature six months later.

In addition to the chatbot, Gemini is powering features that perform complex tasks for users, like preparing research reports and generating podcasts based on a set of documents.

Google sells the technology that powers Gemini to developers, uses it for the “AI overviews” that sometimes appear at the top of search results and recently began packaging it with the business versions of Gmail and Google Drive. Pichai has said all of Google’s products with more than two billion users now have features powered by Gemini.

The company has also tried to convince device manufacturers to make Gemini the default assistant on mobile phones, which it already does on its own Pixel phones. Motorola and OnePlus, which both use Google’s Android mobile software, made Gemini the default on some of their phones last year.

Chinese firms continue to release AI models that rival the capabilities of systems developed by OpenAI and other U.S.-based AI companies.

This week, MiniMax, an Alibaba- and Tencent-backed startup that has raised around $850 million in venture capital and is valued at more than $2.5 billion, debuted three new models: MiniMax-Text-01, MiniMax-VL-01, and T2A-01-HD. MiniMax-Text-01 is a text-only model, while MiniMax-VL-01 can understand both images and text. T2A-01-HD, meanwhile, generates audio — specifically speech.

MiniMax claims that MiniMax-Text-01, which is 456 billion parameters in size, performs better than models such as Google’s recently unveiled Gemini 2.0 Flash on benchmarks like MMLU and SimpleQA, which measure the ability of a model to answer math problems and fact-based questions. Parameters roughly correspond to a model’s problem-solving skills, and models with more parameters generally perform better than those with fewer parameters.

As for MiniMax-VL-01, MiniMax says that it rivals Anthropic’s Claude 3.5 Sonnet on evaluations that require multimodal understanding, like ChartQA, which tasks models with answering graph- and diagram-related queries (e.g., “What is the peak value of the orange line in this graph?”). Granted, MiniMax-VL-01 doesn’t quite best Gemini 2.0 Flash on many of these tests. OpenAI’s GPT-4o and an open model called InternVL2.5 beat it on several as well.

Of note, MiniMax-Text-01 has an extremely large context window. A model’s context, or context window, refers to input (e.g., text) that a model considers before generating output (additional text). With a context window of 4 million tokens, MiniMax-Text-01 can analyze around 3 million words in one go — or just over five copies of “War and Peace.”

For context (no pun intended), MiniMax-Text-01’s context window is roughly 31 times the size of GPT-4o’s and Llama 3.1’s.

The last of MiniMax’s models released this week, T2A-01-HD, is an audio generator optimized for speech. T2A-01-HD can generate a synthetic voice with adjustable cadence, tone, and tenor in around 17 different languages, including English and Chinese, and clone a voice from just 10 seconds of an audio recording.

MiniMax didn’t publish benchmark results comparing T2A-01-HD to other audio-generating models. But to this reporter’s ear, T2A-01-HD’s outputs sound on par with audio models from Meta and startups like PlayAI.

With the exception of T2A-01-HD, which is exclusively available through MiniMax’s API and Hailuo AI platform, MiniMax’s new models can be downloaded from GitHub and the AI dev platform Hugging Face.

Just because the models are “openly” available doesn’t mean they aren’t locked down in certain aspects, however. MiniMax-Text-01 and MiniMax-VL-01 aren’t truly open source in the sense that MiniMax hasn’t released the components (e.g., training data) needed to re-create them from scratch. Moreover, they’re under MiniMax’s restrictive license, which prohibits developers from using the models to improve rival AI models and requires that platforms with more than 100 million monthly active users request a special license from MiniMax.

MiniMax was founded in 2021 by former employees of SenseTime, one of China’s largest AI firms. The company’s projects include apps like Talkie, an AI-powered role-playing platform along the lines of Character AI, and text-to-video models that MiniMax has released in Hailuo.

Some of MiniMax’s products have become the subject of minor controversy.

Talkie, which was pulled from Apple’s App Store in December for unspecified “technical” reasons, features AI avatars of public figures, including Donald Trump, Taylor Swift, Elon Musk, and LeBron James, none of whom appear to have consented to being featured in the app.

In December, Broadcast magazine reported that MiniMax’s video generators can reproduce the logos of British television channels, suggesting that MiniMax’s models were trained on content from those channels. And MiniMax is reportedly being sued by iQiyi, a Chinese video streaming service that alleges MiniMax illicitly trained on iQiyi’s copyrighted recordings.

MiniMax’s new models arrive days after the outgoing Biden administration proposed harsher export rules and restrictions on AI technologies for Chinese ventures. Companies in China were already prevented from buying advanced AI chips, but if the new rules go into effect as written, companies will be faced with stricter caps on both the semiconductor tech and models needed to bootstrap sophisticated AI systems.

On Wednesday, the Biden administration announced additional measures focused on keeping sophisticated chips out of China. Chip foundries and packaging companies that want to export certain chips will be subjected to broader license requirements unless they exercise greater scrutiny and due diligence to prevent their products from reaching Chinese clients.