Benchmark Almanac

Alibaba International Digital Commerce (AIDC-AI)

We track 3 benchmarks that Alibaba International Digital Commerce (AIDC-AI) made or helped make. Its models are on its lab page.

1 in use 2 with public code 0 with a leaderboard here

Benchmarks

Name Kind Released Best score Status
MMMB Models answer multiple-choice questions about images in six languages across 15 categories. The set has about 12,000 questions. Knowledge in many languages Jun 2024 83% In use
Multilingual MMBench Models answer MMBench dev image questions in English, Chinese, Portuguese, Arabic, Turkish and Russian. Scored as accuracy per language. Knowledge in many languages Aug 2024 System cards only
HalloMTBench Models translate English text into one of 11 languages. Judges count translations that add unsupported material, use the wrong language, leave text untranslated or repeat material. Translation Oct 2025 Not in cards