PonyWorld 2.0 adds self-diagnosis to Pony.ai’s L4 stack

Pony.ai has launched PonyWorld 2.0, an upgrade to its proprietary world model that enables its autonomous driving system to identify its own performance gaps, direct targeted data collection, and focus training on the most challenging scenarios. The company says the development marks a step toward a more self-improving approach to level 4 (L4) autonomous driving development.

The system introduces a structured intention layer that allows the model to form an internal representation of its decisions, compare intended outcomes with actual results, and flag scenario types requiring further training. It then generates data-collection tasks for human teams, who gather real-world samples and feed them back into a cloud-based training loop.

Pony.ai said PonyWorld 2.0 is already being applied across its L4 driverless fleet to improve safety, ride comfort and traffic efficiency. The company is targeting a fleet of more than 3,000 vehicles across 20 cities globally by the end of 2026, with nearly half of those deployments in international markets.

In a statement, Dr Tiancheng Lou, Founder and Chief Technology Officer of Pony.ai, said: “PonyWorld 2.0 is an important step toward a more self-improving approach to autonomous driving development. As AI systems become more capable, they can play a larger role not only in learning to drive, but also in guiding their own improvement, making L4 development more scalable over time.”

Source: Pony.ai