Despite the surge in downloads driven by Google's latest model Gemini 3, OpenAI's ChatGPT maintains overwhelming dominance in user engagement and daily habits. The AI race is shifting from acquiring new users to capturing user time—the latter determining which product becomes an ingrained daily habit.
Data from Sensor Tower reveals that in late November, ChatGPT accounted for nearly 90% of mobile chatbot conversations, while Gemini captured only about 4%. Meanwhile, ChatGPT users opened the app over eight times daily on average, far surpassing Gemini's 2.5 times.
OpenAI CEO Sam Altman previously declared a "code red" status, accelerating the release of GPT-5.2 to counter competition. Reports suggest the model could launch as early as December 9, two weeks ahead of schedule.
Google's temporary lead in downloads highlights a critical challenge: attracting users and cultivating habitual usage are distinct battles in the AI market, with the latter being far more difficult.
**Download Share Fluctuates, Usage Depth Vastly Divergent** Following Gemini 3's November 18 launch, Google gained ground in mobile downloads. By the week ending November 24, Gemini's weekly download share rose to 33%, up from 29% in prior weeks, while ChatGPT's share dipped from 45-46% to 42%.
However, downloads didn’t translate into usage. Throughout late November, ChatGPT held nearly 90% of mobile chatbot sessions, with Gemini at just 4%. Perplexity and DeepSeek trailed further. This excludes web traffic.
The gap in engagement frequency was even starker: ChatGPT users averaged over eight daily opens versus Gemini's 2.5, underscoring Google's struggle to convert interest into habitual use.
**OpenAI Addresses Usage Decline Concerns** Under competitive pressure, OpenAI is working to sustain user retention. CFO Sarah Friar recently explained to investors that ChatGPT’s declining usage duration partly stemmed from content restrictions, such as blocking suicide-related discussions with minors.
The firm is rolling out age verification to allow verified adults to discuss mature topics like erotica, balancing safety with experience.
Internally, Altman claimed GPT-5.2 would "outperform Gemini 3 in reasoning." However, analysts note OpenAI’s timelines often shift due to development hurdles, server capacity, or rival moves, potentially delaying the December 9 target.
**Model Performance Fuels Market Anticipation** Unverified social media comparisons suggest GPT-5.2 surpasses Gemini 3 and Claude 4.5 across benchmarks, reflecting high expectations. Reports indicate GPT-5.2 is ready for release as early as December 9, significantly ahead of its late-December schedule—a move highlighting OpenAI’s urgency to defend its lead against Google and Anthropic.
Comments