Google launched its deepest AI research agent yet — on the same day OpenAI dropped GPT-5.2 (techcrunch.com)

Google has unveiled its most advanced AI research agent, Gemini Deep Research, coinciding with OpenAI's release of GPT-5.2. This new agent not only generates detailed research reports but also integrates Google’s SATA-model capabilities into applications through their Interactions API, offering developers enhanced control in the evolving landscape of agentic AI. Google emphasizes that Deep Research can process extensive context and is already being utilized for complex tasks, such as due diligence and drug safety research. The significance of this launch lies in its potential to change how users interact with information—transitioning from conventional search to reliance on AI for comprehensive analysis. Google’s model seeks to minimize hallucinations—errors where AI generates incorrect information—which is especially critical in long-form tasks requiring multiple decisions. To validate its capabilities, Google introduced the DeepSearchQA benchmark for evaluating agents on intricate, multi-step queries, where its agent demonstrated superiority over competitors, though OpenAI’s ChatGPT 5 Pro closely followed. The strategic timing of this announcement, right before OpenAI launched its latest model, highlights the competitive dynamics in the AI landscape as both companies vie for leadership in AI technology.