Google's SVP (responsible for Research, Labs, and Technology & Society) stated in a 2025 research summary that the trajectory of AI has shifted "from tools to utilities." While positioning 2024 as the year the foundation for multimodality was built, the SVP characterized 2025 as the year AI truly began to think, act, and explore the world.
Regarding models, Google released Gemini 2.5 in March, followed by Gemini 3 in November and Gemini 3 Flash in December. The company reported leap-forward progress in reasoning capabilities, multimodal understanding, model efficiency, and generative capabilities. Furthermore, the SVP stated that progress was made in the path toward practical applications of quantum computing, accelerating the commercialization of research.
Source: Google's year in review: areas with research breakthroughs in 2025（HN 245pt・168コメント）（HN Search (backfill）