# AA Global-MMLU-Lite leaderboard

Multilingual benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/multilingual-aa-global-mmlu-lite

Artificial Analysis Global-MMLU-Lite. Every model the catalog carries a published AA Global-MMLU-Lite value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | Multilingual |
| Measure | Accuracy |
| Tasks | Multilingual knowledge questions |
| Difficulty | Multilingual professional knowledge |
| Published by | [Artificial Analysis Global-MMLU-Lite Benchmark Leaderboard](https://artificialanalysis.ai/evaluations/global-mmlu-lite) |

Global MMLU Lite is a multilingual version of MMLU with 400 questions per language, half culturally sensitive and half culturally agnostic. This page holds two independent runs: Artificial Analysis's, in the main table, and Kaggle's.

## AA Global-MMLU-Lite ranking

17 models with a published AA Global-MMLU-Lite value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Accuracy | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [Gemini 3.1 Pro](https://modelscale.dev/models/gemini-3-1-pro) | Google | 93.2 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Gemini 3 Flash](https://modelscale.dev/models/gemini-3-flash) | Google | 92.7 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Claude Opus 4.6 (Adaptive)](https://modelscale.dev/models/claude-opus-4-6-thinking) | Anthropic | 92.2 | Observed 2026-11-21 · source benchlm:models |
| 3 | [Gemini 3 Pro](https://modelscale.dev/models/gemini-3-pro) | Google | 92.2 | Observed 2026-11-21 · source benchlm:models |
| 5 | [Claude Opus 4.5 Thinking](https://modelscale.dev/models/claude-opus-4-5-thinking) | Anthropic | 91.3 | Observed 2026-11-21 · source benchlm:models |
| 6 | [GPT-5](https://modelscale.dev/models/gpt-5) | OpenAI | 90.7 | Observed 2026-11-21 · source benchlm:models |
| 6 | [GPT-5 (high)](https://modelscale.dev/models/gpt-5-high) | OpenAI | 90.7 | Last good 2026-11-18 · source benchlm:models |
| 8 | [GPT-5.1](https://modelscale.dev/models/gpt-5-1) | OpenAI | 90.6 | Observed 2026-11-21 · source benchlm:models |
| 9 | [Gemini 2.5 Pro](https://modelscale.dev/models/gemini-2-5-pro) | Google | 90.3 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Qwen3.5 397B](https://modelscale.dev/models/qwen3-5-397b) | Alibaba | 90.0 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Grok 4](https://modelscale.dev/models/grok-4) | xAI | 89.5 | Observed 2026-11-21 · source benchlm:models |
| 12 | [GPT-5.4 mini](https://modelscale.dev/models/gpt-5-4-mini) | OpenAI | 89.4 | Observed 2026-11-21 · source benchlm:models |
| 13 | [Qwen3.5-122B-A10B](https://modelscale.dev/models/qwen3-5-122b-a10b) | Alibaba | 89.1 | Observed 2026-11-21 · source benchlm:models |
| 14 | [Qwen3.5-27B](https://modelscale.dev/models/qwen3-5-27b) | Alibaba | 86.8 | Observed 2026-11-21 · source benchlm:models |
| 15 | [Qwen3.5-35B-A3B](https://modelscale.dev/models/qwen3-5-35b-a3b) | Alibaba | 86.4 | Observed 2026-11-21 · source benchlm:models |
| 16 | [DeepSeek-R1](https://modelscale.dev/models/deepseek-r1) | DeepSeek | 86.0 | Observed 2026-11-21 · source benchlm:models |
| 17 | [GPT-OSS 120B](https://modelscale.dev/models/gpt-oss-120b) | OpenAI | 82.8 | Last good 2026-09-22 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
