Skip to main content
Meo

MMLU-Pro — AI model leaderboard

AI models ranked by MMLU-Pro, an aggregated third-party benchmark from artificial_analysis. Higher is better. Cross-referenced against our first-party meo scores and Effective Value (𝕍).

Ranking 37 models across the full field · as of 2026-06-07.

#ModelLabMMLU-Pro
1Google: Gemini 3 Flash Previewgoogle88.2%
2OpenAI: GPT-5.2 Chatopenai87.4%
3OpenAI: GPT-5openai87.1%
4OpenAI: GPT-5.1openai87.0%
5OpenAI: GPT-5 Codexopenai86.5%
6Google: Gemini 2.5 Progoogle86.2%
7OpenAI: GPT-5.1-Codexopenai86.0%
8OpenAI: o3openai85.3%
9OpenAI: o1openai84.1%
10OpenAI: GPT-5 Miniopenai83.7%
11OpenAI: o4 Miniopenai83.2%
12Prime Intellect: INTELLECT-3prime-intellect82.2%
13OpenAI: GPT-5.1-Codex-Miniopenai82.0%
14Google: Gemini 2.5 Flashgoogle80.9%
15Meta: Llama 4 Maverickmeta-llama80.9%
16OpenAI: gpt-oss-120bopenai80.8%
17OpenAI: GPT-4.1openai80.6%
18OpenAI: o3 Mini Highopenai80.2%
19Google: Gemini 2.5 Flash Lite Preview 09-2025google79.6%
20OpenAI: o3 Miniopenai79.1%
21OpenAI: GPT-4.1 Miniopenai78.1%
22OpenAI: GPT-5 Nanoopenai78.0%
23Meta: Llama 4 Scoutmeta-llama75.2%
24OpenAI: gpt-oss-20bopenai74.8%
25OpenAI: GPT-4oopenai74.8%
26Xiaomi: MiMo-V2-Flashxiaomi74.4%
27OpenAI: GPT-4o (2024-05-13)openai74.0%
28Microsoft: Phi 4microsoft71.4%
29Cohere: Command Acohere71.2%
30OpenAI: GPT-4 Turboopenai69.4%
31Google: Gemma 3 27Bgoogle66.9%
32Reka Flash 3rekaai66.9%
33OpenAI: GPT-4.1 Nanoopenai65.7%
34OpenAI: GPT-4o-miniopenai64.8%
35Google: Gemma 3 12Bgoogle59.5%
36Microsoft: Phi 4 Mini Instructmicrosoft46.5%
37Google: Gemma 3 4Bgoogle41.7%

Artificial Analysis (artificialanalysis.ai). Redistribution requires an AA commercial license.

← All rankingsMethodology & 𝕍 →