Organisation
01.AI
Formerly a builder of open-weight Yi foundation models; since 2025 refocused on fine-tuning others' open models into enterprise software after stepping back from pretraining from scratch.
01.AI is a Chinese lab founded in 2023 by Kai-Fu Lee, the former head of Google China, which entered the field with the open-weight Yi models — bilingual systems it said matched or beat larger models from Meta and others on standard benchmarks. It was quickly valued at around $1 billion and was, for a time, one of the more prominent Chinese challengers building its own frontier models. In early 2025, however, Lee said pretraining large models from scratch was no longer viable for a startup after DeepSeek's cheap, capable open releases, and repositioned the roughly 200-person company toward fine-tuning others' open models — chiefly DeepSeek's and Alibaba's Qwen — into enterprise software. Its retreat became one of the clearest examples of a well-funded Chinese lab stepping back from the frontier once cheap open-weight models removed the commercial case for training its own.
- Category
- Chinese AI labs
- Founded
- 2023
- HQ
- Beijing, CN
- Key people
- Kai-Fu Lee
Appears alongside
Featured in threads
Tracks
- Open weights & ecosystem 2
- Models & capabilities 2
- Labs & people 1
01.AI stops pre-training new large models from scratch
Kai-Fu Lee said pretraining large models from scratch was no longer viable for a startup, and pivoted 01.AI toward fine-tuning DeepSeek and Qwen for enterprise clients.
Labs & people
DeepSeek releases DeepSeek LLM 67B
DeepSeek's first general-purpose open-weight LLM family, 7B and 67B, trained on 2 trillion English/Chinese tokens.
Open weights & ecosystem · Models & capabilities
01.AI open-sources Yi-6B and Yi-34B
Kai-Fu Lee's 01.AI released its first open-weight models, which it said outperformed larger Llama 2 and Falcon models.
Open weights & ecosystem · Models & capabilities
Also mentioned in 1 entry
Referenced in passing — 01.AI isn't the main subject of these.