# MMLU — AI model rankings

A 57-subject multiple-choice exam — history, law, medicine, maths — that was the standard measure of how much a model knows from 2020 until roughly 2024, when frontier scores crowded into the high 80s and labs moved on to harder tests. The scores here were published years apart under different testing setups, so read them as a historical record rather than a like-for-like ranking. Higher is better.

7 tracked models have a published MMLU score. Higher is better. Scores come from published lab reports and benchmark sources; recorded sources appear beside the scores.

## Ranking

| Rank | Model | Developer | Score | Source | Released |
| --- | --- | --- | --- | --- | --- |
| 1 | gpt-oss-120b | OpenAI | 90% | Lab | Aug 5 2025 |
| 2 | DeepSeek V3 Chat | DeepSeek | 88.5% | Lab | Dec 26 2024 |
| 3 | DeepSeek V3 Base | DeepSeek | 87.1% | Lab | Dec 25 2024 |
| 4 | GPT-4 | OpenAI | 86.4% | Lab | Mar 14 2023 |
| 5 | gpt-oss-20b | OpenAI | 85.3% | Lab | Aug 5 2025 |
| 6 | Mistral Small 3.1 | Mistral | 80.62% | Lab | Mar 17 2025 |
| 7 | Claude 2 | Anthropic | 78.5% | Lab | Jul 11 2023 |


---

Canonical page: https://aireleasetracker.com/benchmark/mmlu
Site index: https://aireleasetracker.com/llms.txt
Source: AI Release Tracker (https://aireleasetracker.com). Most benchmark scores come from lab launch material; gathered results identify the leaderboard that published them.