Jingqi Tong

Even the best large vision-language models struggle with multi-image reasoning, scoring only 50% on a new benchmark designed to challenge their capabilities.

Forget benchmarks: AI can now learn "scientific taste" and propose research ideas with higher potential impact than humans, thanks to a novel reinforcement learning approach using citation data.

GPT-5's scientific reasoning skills plummet by nearly 50% when tackling multi-step workflows, revealing a critical gap in current LLM agents' ability to orchestrate complex tool use.

Open-source MOVA lets you generate synchronized, high-quality video and audio—including realistic lip sync—without relying on closed-source systems.