Two new research papers introduce advanced agent systems designed for deep research tasks. The first, QUEST, offers a family of open-weight models (2B to 35B parameters) trained on synthetic data, demonstrating strong performance in fact-seeking, citation grounding, and report synthesis, rivaling proprietary agents. The second, Argus, presents a cooperative Searcher-Navigator system built on a 35B MoE backbone, which excels at assembling evidence from complementary sources, achieving state-of-the-art results on benchmarks like BrowseComp while maintaining a manageable context window. AI
IMPACT These advancements in open-weight deep research agents and scalable evidence assembly could accelerate knowledge synthesis and democratize access to advanced AI research capabilities.
RANK_REASON Two distinct research papers introducing new agent architectures and training methodologies for deep research tasks.
Read on arXiv cs.IR (Information Retrieval) →
AI-generated summary · Google Gemini · from 4 sources. How we write summaries →
