ARC-3, a sneak peek at the next-gen, interactive reasoning benchmark designed to illuminate the capability gap between today's AI and tomorrow's AGI.
Play First 3 Games
https://three.arcprize.org/
As with previous ARC tests, the actual games used for testing AI are kept secret. AI algorithms must learn the games on the spot.
There are no instructions. You must play the game to discover controls, rules, and goal.
Interactive Reasoning Benchmarks (IRBs) test for a broad scope of capabilities:
• Exploration
• Percept -> Plan → Action
• Memory
• Goal Acquisition
• Alignment
Game Design Constraints
• Easy for humans (can pick it up in <1 min of game play)
• Core Knowledge Priors (no language, trivia, cultural symbols)
• Should require no instructions to play
• Should be fun for humans and playable in 5-10 minutes
• Innovative and novel game mechanics encouraged (Hidden state, theory of mind, long term planning, navigating other agents, etc.)