Google announced this week that its Gemini app has surpassed 1 billion monthly users, even as independent benchmark testing places the company's strongest models behind those of competitors Anthropic and OpenAI.

Gemini Reaches 1 Billion Monthly Users

The announcement came during Google's ongoing Made by Google showcase. The company described Gemini as the fastest-growing product in its history and signaled it is now pursuing the next billion users.

Gemini benefits from deep integration across Google's existing product ecosystem, including Android, Workspace, and Search, giving it distribution channels that competitors operating primarily through standalone applications do not have.

Voice interaction has become a primary driver of engagement. According to Google, 63% of Gemini users communicate with the assistant by voice, with a subset relying on voice exclusively. The company noted that busy parents are 43% more likely to use voice for everyday tasks. Image generation has also scaled significantly, with the app producing more than 150 million images per day.

On Apple's iOS, Gemini has surpassed 100 million active users. macOS users rank among the most active, submitting prompts at approximately double the rate of users on other platforms. On Android, Gemini can now trigger in-app actions across more than 40 applications.

New Models Ship Ahead of Milestone

Days before the user milestone, Google released three new models focused on cost efficiency and scale: Gemini 3.6 Flash, Gemini 3.5 Flash-Lite, and Gemini 3.5 Flash Cyber.

Gemini 3.6 Flash now serves as Google's default production model, operating approximately 17% leaner on output tokens and supporting a context window exceeding one million tokens.

Flash Cyber marks Google's first Gemini model purpose-built to identify and patch software vulnerabilities. Access is currently restricted to governments and vetted partners under a limited program.

The emphasis on Flash-tier efficiency comes as AI providers across the industry face growing pressure to reduce inference costs while maintaining output quality, particularly for high-volume consumer applications.

Benchmark Performance Lags Competitors

On the Artificial Analysis Intelligence Index, Anthropic's Claude Opus 5 leads with a score of 63, while Google's highest-placed entry, Gemini 3.6 Flash, scores 52. Models from OpenAI, xAI, and Kimi also rank above or level with Google's top performer.

Artificial Analysis does rate Gemini 2.5 Flash-Lite as the fastest model it tracks, with a time-to-first-response of 0.31 seconds. However, the model trails on composite intelligence benchmarks that evaluate coding, reasoning, and knowledge collectively.

The gap between Gemini's user adoption and its benchmark standing underscores a broader industry pattern: consumer reach in generative AI is shaped not only by model quality rankings but also by distribution, product integration, and brand familiarity.

Expanded Integrations and Accessibility

Google announced during the showcase that Gemini will gain connections to additional services over the coming weeks, including Wix, OpenTable, Ticketmaster, Pandora, Angi, and Zocdoc.

Google DeepMind also introduced SL2T, a sign-language-to-text model trained on more than 100,000 hours of data spanning over 50 sign languages. The model powers sign-to-text dictation in Gboard and Live Transcribe on the Pixel 11, launching first with American Sign Language to English. The feature enables Deaf users to sign to their phones in contexts where typing would normally be used, as detailed on DeepMind's blog.