Yi
01.AI · November 2023
● activeOpen Sourcedecoder onlytext
Parameters6B - 34B
Context Window200K tokens
Variants6B, 34B
Why It Matters
One of the first models to demonstrate that Chinese AI startups could build competitive open-source models rivaling Western labs.
Description
A strong bilingual (English/Chinese) model family from 01.AI, the startup founded by renowned AI researcher Kai-Fu Lee. Available in 6B and 34B sizes with an impressive 200K token context window (roughly 150,000 words), it was one of the first models to demonstrate that Chinese AI startups could build competitive open-source models rivaling those from major Western labs.
Key Innovations
Long Context
Long ContextAbility to process very long inputs (100K+ tokens), enabling analysis of entire codebases or books.
Open Weight
Open WeightModel weights are publicly released but training data/code may not be. Enables fine-tuning but not full reproduction.