AI OrbitExplore • Compare • Stay Ahead
VERIFIED AI BENCHMARK

MMLU Leaderboard

Explore verified MMLU AI benchmark results, rankings, methodology, sources and tested models on AI Orbit.

Knowledge · % · Higher is better · Version 2025 · standard
Only verified results are included in this public leaderboard.

Official methodology / benchmark source

#Model / ToolScoreTestedSource
1GPT-4o85.70 %Apr 14, 2025OpenAI GPT-4.1 announcement
2GPT-4o mini82.00 %Apr 14, 2025OpenAI GPT-4.1 announcement

About MMLU

This leaderboard uses the latest verified result available for each listed entity. Results retain their source and test date so readers can evaluate freshness and methodology.