Browse Models
The Yi model family, developed by 01.AI, comprises bilingual large language models ranging from 6 billion to 34 billion parameters, designed for English and Chinese language tasks. Built on decoder-only Transformer architectures with features like Grouped-Query Attention and extended context windows up to 200,000 tokens, the models include base, chat-optimized, and vision-language variants released under Apache 2.0 licensing for research and commercial use.