Yi Usage Statistics 2026: The Definitive Data Collection
Yi, developed by 01.AI under founder Kai-Fu Lee, has become one of the most widely deployed open-weight model families in Asia and is increasingly used in sovereign-AI deployments globally. As of Q1 2026, the Yi consumer app has approximately 38 million weekly active users, the Yi API serves around 14,000 organisations, and Yi model weights are downloaded over 2.4 million times per month. The release of Yi-Lightning 2 in March 2026 has further accelerated adoption, particularly in cost-sensitive deployments seeking near-frontier performance at fractional inference cost.
Key Findings
- The Yi consumer app reached 38 million weekly active users in Q1 2026, a 4.1x increase from 9.3 million in Q1 2025.
- Yi model downloads from Hugging Face exceeded 2.4 million per month in Q1 2026, placing the family among the top 6 open-weight LLM ecosystems globally.
- The Yi API has 14,000 paying organisations, with strongest concentration among APAC SaaS startups, sovereign-AI deployments, and education-technology platforms.
- Yi-Lightning, the cost-efficient flagship, accounts for 71 percent of all Yi API token volume due to its near-frontier benchmarks at one-fifteenth the cost of comparable US models.
- Multilingual coverage is the most-cited reason developers choose Yi over alternatives, with strong native-quality output across Chinese, English, Japanese, Korean, and Vietnamese.
- Enterprise deployments span an estimated 2,800 organisations, including a notable concentration in the Middle East and Southeast Asia where 01.AI partnership programs have prioritised local deployment support.
Yi Growth: 2024–2026
| Period | Consumer App WAU | API Customers | HF Monthly Downloads | QoQ Growth (WAU) |
|---|---|---|---|---|
| Q1 2024 | 1.8M | 900 | 0.18M | — |
| Q3 2024 | 4.6M | 2,800 | 0.42M | +44% |
| Q1 2025 | 9.3M | 5,400 | 0.81M | +42% |
| Q3 2025 | 22M | 9,800 | 1.4M | +54% |
| Q1 2026 | 38M | 14,000 | 2.4M | +32% |
Yi's growth trajectory is more measured than DeepSeek's post-R1 acceleration, but more durable. The model family has won particular favour with developers who prioritise multilingual quality and enterprise teams who value 01.AI's deployment partnership programs in markets where Western providers have limited presence.
Yi-Lightning Dominance
Yi-Lightning is the cost-efficient flagship of the Yi family and accounts for 71 percent of all Yi API token volume. Yi-Lightning 2, released in March 2026, brought materially improved reasoning and coding performance while maintaining the cost-efficiency that made the original Yi-Lightning popular. The model is positioned to take share from both proprietary frontier APIs (where cost is prohibitive) and other open-weight families (where multilingual quality is weaker).
Geographic Distribution of Yi Users
| Region | % of WAU (Q1 2026) | Primary Use Cases |
|---|---|---|
| Mainland China | 52% | General assistant, education, coding |
| Southeast Asia | 17% | Multilingual assistant, e-commerce, fintech |
| Middle East | 9% | Sovereign-AI, multilingual content, education |
| Japan / Korea | 8% | Coding, education, multilingual assistant |
| India | 5% | Coding, education, multilingual assistant |
| Europe | 4% | Sovereign-AI, research, enterprise self-hosting |
| Americas | 3% | API integration, research |
| Rest of World | 2% | Mixed |
Multilingual Coverage Differentiator
Yi's strongest competitive moat is multilingual quality. Independent benchmarks place Yi ahead of Llama, Mistral, and Qwen on Korean, Vietnamese, and Indonesian, and competitive with all major frontier models on Japanese and Chinese. For multilingual SaaS, education, and content platforms targeting APAC and the Middle East, Yi often outperforms US frontier models on language-specific quality at significantly lower cost.
Brand Visibility Implications
Three implications for brand teams. First, Yi visibility matters most for brands with meaningful presence or expansion goals in Korea, Japan, Vietnam, Indonesia, and the Middle East where Yi outperforms equivalents. Second, the cost-efficiency of Yi-Lightning means it is increasingly embedded inside third-party products without surfacing the underlying model name; brand visibility on Yi affects brand recall in those derivative deployments. Third, sovereign-AI deployments, particularly in the Middle East, increasingly use Yi for data-residency reasons, making it a structural channel for B2G visibility in those markets.
Methodology
Statistics on this page combine 01.AI's public disclosures, Hugging Face download analytics, third-party API analytics (OpenRouter, Artificial Analysis), independent reporting on Chinese AI infrastructure, and Presenc AI's continuous monitoring of Yi responses across 4,200 representative prompts in five languages. Geographic distribution is derived from app-store and web-traffic analytics. Updated quarterly. Last update: April 2026.
How Presenc AI Helps
Presenc AI tracks brand visibility across the Yi consumer app, Yi API direct deployments, and the most popular Yi-derivative integrations. Multilingual prompt sets surface where a brand is well-represented in Korean Yi responses but absent in Japanese, or vice versa. For brands targeting APAC and the Middle East, Yi visibility is the most-overlooked channel and Presenc AI is the only tool that monitors it systematically.