AI Leaderboard Shake-Up: Open Models Fall 12 Points Behind #shorts

A new hidden benchmark puts open AI models 12 points behind frontier systems, while Claude Opus 4.6 reportedly uncovered the answer key. Public tests can be trained for—so how long will this one stay reliable?

🎧 Watch the full episode of What the AI?!
YouTube, Spotify, Apple Podcasts → https://www.whattheai.fm

🔗 Watch the full video: https://youtu.be/5JxvSNvXPVA

#AILeaderboards #AIModels #BenchmarkTesting #OpenSourceAI