Loading the catalogue…
🇨🇳China · Mixed · 0 models
01-ai✓
01.AI (零一万物) is a Beijing-incorporated AI company known for the Yi series of bilingual open-weights large language models, all now released under Apache 2.0, making self-hosted EEA deployment feasible without contacting 01.AI infrastructure. However, the company is subject to China's National Intelligence Law (2017), which can compel state cooperation without public disclosure — a structural, non-waivable sovereignty risk for EU regulated customers that no contractual instrument can mitigate. Since March 2025, 01.AI has halted pre-training its own foundation models and pivoted to reselling enterprise solutions built on DeepSeek's models, placing the Yi series in effective maintenance mode and raising material uncertainty about continued security patches and long-term support; the company publishes no GDPR-compliant privacy policy, DPA, or EU AI Act compliance documentation.
AI Act 47/100Outside CLOUD ActHF verified
Stav AI Act verdict
jurisdictionhigh
As a PRC-incorporated entity, 01.AI is subject to China's National Intelligence Law (2017), which can compel cooperation with Chinese state intelligence services without public disclosure. This creates a structural, non-waivable data sovereignty risk for EU regulated enterprises that is analogous in severity — though legally distinct in mechanism — to US CLOUD Act exposure. No contractual instrument can mitigate this obligation.
privacyhigh
No GDPR-compliant privacy policy, Data Processing Agreement, or Data Protection Officer contact is publicly available in English. The official website is a Chinese-language Sxl.cn/Strikingly site with no EU-facing compliance documentation. EU enterprises cannot establish a lawful basis for any data sharing with 01.AI under GDPR Chapter V without these instruments.
stabilitymedium
As of March 2025, 01.AI has halted pre-training of large language models and pivoted to reselling enterprise AI solutions built on DeepSeek's models. The Yi series is in effective maintenance mode, raising material uncertainty about long-term security patches, model updates, and ongoing support commitments.
compliancemedium
No EU AI Act compliance statement, GPAI transparency documentation (Article 53), or GPAI Code of Practice participation found. As a non-EU company distributing models globally, 01.AI may be in scope for EU AI Act obligations without having taken steps to comply.
securitymedium
No security certifications (SOC 2, ISO 27001), bug bounty programme, or published vulnerability disclosure policy identified. EU regulated customers cannot independently verify the security posture of 01.AI's model development and distribution pipeline.
governancemedium
Corporate governance transparency is limited. 01.AI has not published a clear board structure, independent oversight mechanisms, or corporate governance documentation beyond founder-CEO Kai-Fu Lee's public profile. The March 2025 pivot and January 2025 denial of acquisition rumours suggest strategic uncertainty at the leadership level.
governancelow
Yi-34B's initial release (November 2023) obscured its architectural basis in Meta's Llama 2 by renaming tensors without attribution. While subsequently acknowledged and corrected — and assessed as technically within normal practice by EleutherAI — this episode reflects a historical transparency lapse. The GitHub README now explicitly credits Llama architecture.
Stav AI Act assessment
Editorial assessment, not legal advice. Stav's risk ratings, scores, and verdicts are our own analysis of publicly available information and may be incomplete or out of date. Verify independently before making compliance or procurement decisions.
Safeguards in place
- complianceUnknown
All current Yi series model releases (Yi base, Yi-1.5, Yi-Coder) are distributed under Apache 2.0 — a permissive, commercially friendly licence with no PRC-law compliance requirements. This resolves the key licence risk flagged at the November 2023 launch and enables EEA self-hosting without restriction.
- transparencyUnknown
Yi technical report published on arXiv (arXiv: 2403.04652) and model cards available on GitHub with architectural details, benchmark results, and training data descriptions, providing meaningful technical transparency for self-hosted model evaluation.
- transparencyUnknown
Following the November 2023 Llama attribution controversy, 01.AI publicly acknowledged the oversight and the GitHub README now explicitly credits the Llama architecture and community, demonstrating responsiveness to community accountability.
- stabilityUnknown
Backed by major Chinese technology companies (Alibaba Group, Xiaomi, Tencent) and Sinovation Ventures, with a second round from international and Southeast Asian investors in August 2024, providing financial resilience for an organisation no longer investing in expensive pre-training compute.
- communityUnknown
Yi-34B achieved first place on the HuggingFace Open LLM Leaderboard (pre-trained) and C-Eval benchmarks at launch (November 2023), and Yi-Lightning ranked joint third among LLM providers in October 2024 (UC Berkeley/LMSYS ranking), demonstrating genuine technical capability.
- securityUnknown
01.AI's GitHub README notes use of 'data compliance checking algorithms during the training process' — a publicly stated, if minimal, training-data quality control measure, providing a baseline level of disclosed practice.
Published safeguards & certifications
Privacy posture
- Flagged issues
- 1
- No GDPR-compliant privacy policy, Data Processing Agreement, or Data Protection Officer contact is publicly available in English. The official website is a Chinese-language Sxl.cn/Strikingly site with no EU-facing compliance documentation. EU enterprises cannot establish a lawful basis for any data sharing with 01.AI under GDPR Chapter V without these instruments.high
Privacy policy review
Identity & profile
- Country
- 🇨🇳 China
- Open-weights policy
- Mixed
- Models in catalogue
- 0
- Website
- https://www.01.ai
- GitHub
- https://github.com/01-ai
- HuggingFace
- 01-ai · Verified · 834 followers
- HF published
- 28 models · 0 datasets · 0 papers
Creator profile
About 01-ai
01.AI (零一万物) is a Beijing-based AI company founded by Kai-Fu Lee, focused on large language model development and, since early 2025, on enterprise AI solutions built on third-party foundation models. The company releases its Yi series under an open-weights but custom, restrictive licence that explicitly subjects users to Chinese law, raising significant concerns for EU-regulated deployments. There is no published GDPR compliance posture, EU AI Act statement, or independent security certification, and the company's strategic pivot away from pre-training — now reselling DeepSeek-based solutions — raises questions about its long-term model maintenance commitments.
Stav editorial summary
COMPLIANCE & TRUST
Stav’s creator score.
Stav compliance has not yet scored 01-ai. Scores are published once the policy review and infrastructure assessment complete.
Findings
Findings from privacy & terms
Citations gathered when the Compliance Curator last reviewed this creator’s public-facing documents. Grouped by source so the picture stays auditable.
en.wikipedia.org · 2 citations
“01.AI (Chinese: 零一万物; pinyin: Língyī Wànwù) is an artificial intelligence (AI) company based in Beijing, China. ”
“In November 2023 the company released its first open‑source model, Yi-34B. ”
blog.eleuther.ai · 1 citation
“So, even though Yi-34B adopts Llama 2's architecture, Meta's model did not give 01.AI access to any previously inaccessible innovation. In November 20...”
Editorial sources
- 01.AI - Wikipedia
01.AI (Chinese: 零一万物; pinyin: Língyī Wànwù) is an artificial intelligence (AI) company based in Beijing, China.
- Valued at $1B, Kai-Fu Lee's LLM startup unveils open source model | TechCrunch
Its vice president of technology, for instance, is an early member of Google’s Bard, and its chief architect was a founding member of TensorFlow and w...
- Meet China's top six AI unicorns: who are leading the wave of AI in China · TechNode
In the first three quarters of this year, 01.AI launched two models—Yi-Lightning and Yi-Large—both of which ranked among the top on LMSYS and were the...
- Chinese tech unicorn 01.AI admits ‘oversight’ in changing name of AI model built on Meta Platforms’ Llama system | South China Morning Post
Photo: SCMP · “During extensive training experiments, we made several renamings in the code to meet experimental requirements,” 01.AI’s Lin said in hi...
- Kai-Fu Lee - Wikipedia
He has served as co‑chair of the World Economic Forum’s Global AI Council and was named to TIME's 2023 list of the 100 most influential people in AI.
Reference citations
- 01.AI - Wikipedia
01.AI (Chinese: 零一万物; pinyin: Língyī Wànwù) is an artificial intelligence (AI) company based in Beijing, China.
- Kai-Fu Lee calls LLM "historical opportunity" for China as he launches AI startup · TechNode
01.AI was formally founded on May 16, according to corporate database Qichacha, with Ma Jie, former head of Baidu’s metaverse unit, holding a 99% stak...
- Kai-Fu Lee - Wikipedia
In November 2023 the company released its first open‑source model, Yi-34B.
- 01-ai/Yi-34B · License
I was wondering if you had any plans to change the license to a more standard license such as Apache 2.0 or MIT. The Yi License currently pretty restr...
- Just an FYI,the Yi license is a custom license an is not what I would describe a... | Hacker News
It's non-commercial only: "If you plan to use the Yi Series Models and Derivatives for Commercial Purposes, you should contact the Licensor ...
- 01.AI: How Kai-Fu Lee Shapes the AI Narrative in Both the U.S. and China
Shortly after the release of the Yi model series (November 2023), developers spotted that 01.AI's code initially referenced Meta’s model, though these...
- Yi-34B, Llama 2, and common practices in LLM training: a fact check of the New York Times | EleutherAI Blog
So, even though Yi-34B adopts Llama 2's architecture, Meta's model did not give 01.AI access to any previously inaccessible innovation. In November 20...