shawn0728/OpenSearch-VL

[ICML 2026 & EMNLP 2026] Multimodal deep-research MLLM and benchmark. The first long-horizon multimodal deep-research MLLM, extending the number of reasoning turns to dozens and the number of search-engine interactions to hundreds.

DeepDive: Advancing Deep Search Agents with Knowledge Graphs and Multi-Turn RL

Awesome Multimodal Agent

🚀 Ultra Recipe for Training Long-Horizon Search Agents - matching frontier AI's search capability with a 20B model + stateful harness

Search-R1: An Efficient, Scalable RL Training Framework for Reasoning & Search Engine Calling interleaved LLM based on veRL

The Level-Navi Agent, a framework that requires no training and utilizes large language models for deep query understanding and precise search operations. The repo includes benchmarks, datasets, and tools for assessing LLM performance in web searches

Climbing to the Search Frontier.

HyperEyes is a parallel multimodal search agent that fuses visual grounding and retrieval into a single atomic action, enabling concurrent search across multiple entities while treating inference efficiency as a first-class training objective.