Google launches advanced AI research agent Gemini Deep Research alongside OpenAI's GPT-5.2

TechCrunch
TechCrunch
1m ago • 1 views
Google introduced its Gemini Deep Research AI agent with developer API concurrent with OpenAI's GPT-5.2 launch, marking intensified AI rivalry.
Google launches advanced AI research agent Gemini Deep Research alongside OpenAI's GPT-5.2
A What happened
On December 11, 2025, Google released Gemini Deep Research, an AI agent based on its Gemini 3 Pro foundation model. Unlike previous versions, this agent supports developer integration via a new Interactions API. It processes extensive information to perform tasks like due diligence and drug toxicity research, aiming to reduce AI hallucinations during complex multi-step reasoning. Google introduced the DeepSearchQA benchmark to validate performance, where its model led but faced close competition from OpenAI's GPT-5.2 released the same day. OpenAI claims GPT-5.2 surpasses competitors across benchmarks, reinforcing a swift escalation in AI capabilities.
★

Key insights

  • 1

    Rise of agentic AI altering human-information interactions: Google's integration of an API allowing developers to embed advanced AI research agents signals a shift from direct human querying to autonomous AI intermediaries, potentially changing how information is consumed and decisions are made.

  • 2

    Factual accuracy in AI is crucial for multi-step autonomous tasks: Minimizing hallucinations is critical as AI agents undertake longer, complex sequences of decisions where one error can invalidate entire outputs, reflecting a key technical challenge in advancing agentic AI.

  • 3

    Intensifying AI competition drives rapid benchmark innovation and releases: Google and OpenAI's simultaneous releases, each unveiling new benchmarks and claiming leadership, underscore how competitive pressures accelerate AI capability announcements and benchmarking efforts.

Takeaways

Google's Gemini Deep Research and OpenAI's GPT-5.2 launches exemplify intense competition driving rapid progress in agentic AI, shifting how complex research and autonomous tasks are managed.

Topics

Technology & Innovation AI & ML Science & Research Research Business & Markets Innovation