Yi
A bilingual Chinese–English open-weight model with very-long-context versions
WHAT IT IS
Yi is an open-weight model series 01.AI released in November 2023. 01.AI was founded in Beijing in 2023 by Kai-Fu Lee. Positioned as a bilingual Chinese–English general model, Yi ships in sizes such as 6B and 34B, with versions that extend the context to the 200K-token range. It offers both base and instruction-tuned models, with weights downloadable under a permissive licence. Its bilingual ability and long-context options make it a common base for Chinese–English tasks in the open-source community.
Why it matters
At a time when open models leaned English, it made “balanced Chinese–English plus very long context” an explicit selling point and covered different compute budgets with 6B and 34B sizes — a common base choice in the Chinese community.
Key specs
- Parameters
- 6B / 34B (Yi family)
- Context window
- 200K tokens (Yi-34B-200K)
- Open weights
- Yes
- Released
- 2023-11
Capabilities
Related concepts
Transformer Architecture
Replacing word-by-word relay with a room where everyone speaks at once, so long-range dependencies are one hop away
Pretraining & Fine-tuning
Learn language first from vast unlabelled text, then specialise with little data — the most data-efficient paradigm in modern AI
Tokenization
Models do not read characters, they read tokens — and how you split text quietly sets both capability and cost
Comparable products
Qwen
2023An open-weight family spanning many sizes, with multimodal versions
Baichuan
2023An open-weight general model aimed at Chinese-language use
Llama
2023The model family that made the open-weight route mainstream
Mistral Large
2024The flagship commercial model from a European open-weights lab