Skip to content

The Aggregate

Unified LLM rankings, daily

Models What's New Trends Benchmarks Skill Maps
How It Works Guesswork Metrics
The Aggregate
Loading data...

AA Global-MMLU-Lite - Arabic — leaderboard

Metric: Accuracy (%). Source: artificialanalysis.ai. 119 models tracked.

Top models

#ModelScore
1Gemini 3.1 Pro (Preview)93
2Gemini 3 Pro (Preview) (High)92.67
3Claude Opus 4.6 (Adaptive Reasoning, Max Effort)92.25
4Claude Opus 4.5 (Thinking)91.33
5Claude Opus 4.5 (Non-reasoning)90.92
6GPT-5 (High)90.42
7Gemini 2.5 Pro90.33
8Claude Sonnet 4.6 (Adaptive Reasoning, Max Effort)90.17
9GPT-5.2 (Medium)89.83
10GPT-5 (Medium)89.75
11DeepSeek V3.2 Speciale89.58
12GPT-5.1 (High)89.5
13Qwen 3.5 122B A10B (Reasoning)89.5
14Claude Sonnet 4.5 (Thinking)89.42
15Grok 4.20 0309 (Reasoning)88.75

Interactive version: theaggregate.ai/benchmark?slug=aa-global-mmlu-lite-arabic · How the rankings work · Data refreshed daily, snapshot 2026-07-22.

Built by Mikhail Doroshenko — AI researcher, co-author of Humanity’s Last Exam. Independent project; no affiliation with any AI lab.