Google Debuts Gemini 2.5 AI Model, Taking Aim at ChatGPT’s Market Dominance

Google has unveiled Gemini 2.5, its latest experimental artificial intelligence (AI) model, aiming to enhance reasoning capabilities across various domains. Announced on March 25, 2025, Gemini 2.5 Pro is now available for developers to explore. Google describes it as a "thinking model" that processes information step-by-step before delivering responses, thereby improving accuracy and performance, especially in coding, science, and mathematics tasks. The model also supports the development of more context-aware agents.

In competitive assessments, Gemini 2.5 has demonstrated notable performance. It secured the second position on LMArena's WebDev leaderboard, a real-time AI coding competition focused on web development challenges. With an arena score of 1267.70, Gemini 2.5 outperformed models like DeepSeek, Grok, and ChatGPT. However, it was surpassed by Anthropic's Claude 3.7 Sonnet, which achieved a score of 1354.01.

Despite the emergence of new AI models, OpenAI's ChatGPT continues to dominate the AI tools market. In 2024, ChatGPT recorded over 40 billion visits, representing nearly 40% of the total market share. Canva's AI generator followed with a 10.25% share, equating to 10.4 billion visits. More recently, DeepSeek's AI tools have gained traction, securing a 6.58% market share in February 2025, with a growth rate of 195% and 792 million monthly visits. Nevertheless, ChatGPT maintains a significant lead, holding a 43% market share with 5.2 billion monthly visits during the same period.

These developments underscore the rapidly evolving landscape of AI technologies, with companies like Google and Anthropic introducing advanced models to challenge established leaders like ChatGPT. As AI continues to advance, the competition among these models is expected to drive further innovation and improvements in the field.