About Artificial Analysis Global-MMLU-Lite
Definition and scoring
- Organisation
- Artificial Analysis
- Category
- Knowledge
- Version
- 2026
- Direction
- higher is better
- Ranking use
- Reference
- Contamination risk
- Unknown
An independently evaluated multilingual knowledge result from Artificial Analysis. Every genuine source row stays tied to its exact model label, configuration, benchmark version and evaluation system. The summary chart shows one best compatible score per canonical product; Score 2.0 admits only explicitly mapped, frozen protocols and keeps incompatible configurations separate.
Open benchmark source ↗