Build a News Aggregator

Favicon for TechCrunch
Relaxing app background

A new, challenging AGI test stumps most AI models

The Arc Prize Foundation has a new test for AGI that leading AI models from Anthropic, Google, and DeepSeek score poorly on.