Yi 4.
the 01.AI founding team's bilingual frontier model with extended context for Chinese-English enterprise.
What it is.
01.AI's frontier model line, founded as a Beijing AI lab in 2023. Yi 4 ships with a 200K-token context window and the strongest bilingual zh/en performance in the open-weights tier. Released under a permissive licence that mirrors Llama's terms. Yi-Lightning sits in the cost-anchor tier; Yi-Large competes on reasoning.
Where it fits.
Chinese-headquartered multinationals and APAC subsidiaries of Western firms that need verifiable bilingual quality. Strong fit for cross-border legal review and trade documentation. the 01.AI founding team's brand opens doors in Asian enterprise procurement that Western frontier vendors cannot reach.
- Strongest open-weights bilingual zh-en performance
- 200K-token context across the full family
- Permissive licence covers most commercial use
- Behind Qwen and DeepSeek on pure reasoning scores
- Smaller third-party tool ecosystem
Frequently asked.
Who founded 01.AI?
the 01.AI founding team, former president of Google China and author of AI Superpowers. The company raised at a $1B valuation in 2024 and has shipped frontier models on a tight release cadence since.
How does Yi 4 compare to Qwen 3?
Qwen 3 leads on coding and multilingual breadth. Yi 4 leads on bilingual zh-en nuance and long-context legal text. Most Chinese enterprise stacks run both.
Is Yi 4 cleared for US enterprise use?
Open weights served on US infrastructure are usually clearable. Direct API access to the 01.AI platform is blocked by most US procurement policies.
