Tag: #multimodal-agent5 results
A domestically developed multimodal Agent achieves state-of-the-art performance in medical image segmentation without model modifications or extra tokens. The research has been accepted to CVPR 2026, highlighting efficient agent-based approaches.
xAI is rolling out Imagine Agent in Grok Imagine, allowing users to direct an agent to create images and videos. The update includes an open Canvas AI workspace for enhanced creativity.
Researchers from CUHK MMLab, USTC, and Xiaohongshu built a multimodal deep-research LLM for real-world search, tackling image retrieval hit rates and reasoning depth via data synthesis and RL. It performs dozens of reasoning rounds and hundreds of search interactions. The model achieves SOTA on 6 benchmarks with smaller parameters than competitors.
Xiaomi's MiMo large model introduces its first Token Plan, enabling a single subscription to handle full-modality Agent tasks. First-time purchases receive an 88% discount.
Huang Yong launches Qiduo's AI science camera for 6-12 year olds, prioritizing intelligent agent over hardware. Expands '10k Whys' to 1M Q&A dataset for multimodal cognition. Company secures two funding rounds in under a year.