Apex Intelligence

AI models that iterate on research

Apex Intelligence's research system is designed to take a scientific problem through repeated hypothesis generation, experiment design, validation and refinement. Its approach focuses on mid-training and post-training methods that let models improve their reasoning pathways and use tools as they work through a research task.

The Beijing-based company describes itself as a foundation-model research lab, with work spanning large models, reinforcement learning, agentic systems and AI for science. Its technology remains under development; public materials frame scientific discovery as the research objective rather than describing a mature commercial product.