Only three AI models finished above starting capital in a 500-day startup survival test
Researchers at Princeton University built CEO-Bench, a test where AI agents have to run a fictional software company for 500 simulated days. Most current models go broke, and a simple rule-based heuristic with no AI beats nearly all of them.
AI agents are getting increasingly good at narrow tasks: fixing a bug, following a service policy in a conversation, or completing a web-based workflow. These tasks share a simple structure, according to the Princeton study: the agent gets a clear goal, acts briefly, and receives quick feedback. Many important real-world tasks look nothing like that. They involve long chains of decisions under uncertainty, where you have to set priorities, allocate limited resources, read noisy signals, and adapt to changing conditions. To test exactly these skills, the researchers developed CEO-Bench. The benchmark simulates a realistic example of this kind of long-horizon task: running a startup for 500 simulated days. The researchers point to a famous example: in 1997, Apple was 90 days from bankruptcy. Steve Jobs drew a simple two-by-two grid—consumer and pro, desktop and portable—and decided Apple would only build products for those four quadrants. The iMac, iPod, and iPhone followed. This type of strategic steering intelligence is fundamentally different from what AI agents do today, the authors argue. Agents are getting better at individual tasks fast. But steering an entire organization toward long-term goals? That’s a different problem entirely. CEO-Bench is a first attempt at measuring exactly this “steering intelligence.” In CEO-Bench, an agent runs a made-up subscription software company called NovaMind. It starts with zero customers and one million dollars in the bank.
