OpenBMB
To build efficient foundation models and systems that bring large-scale AI capabilities to edge devices and every household.
At a Glance
- Consumer Electronics (Smartphones/IoT)
- Telecommunications
- Enterprise Software
- AI Research Community
AI Tools by OpenBMB
(2)
MiniCPM
Efficient Edge Device LLMs
ChatDev
Multi-Agent AI Software Development
Discussions
No discussions yet
Be the first to start a discussion about OpenBMB
Latest News
ModelBest secures major funding led by China Telecom for on-device AI acceleration.
OpenBMB releases MiniCPM 4.1 and MiniCPM-V 4.5 multimodal flagship models.
ModelBest closes new financing round of hundreds of millions of yuan.
Launch of MiniCPM-V 2.6, providing high-quality multimodal performance for edge devices.
Products & Services
A family of edge-side large language models designed to achieve performance comparable to much larger models like Llama.
Multimodal versions of MiniCPM capable of processing images, videos, and text on mobile devices.
An efficient training toolkit for large-scale models, optimizing memory and compute usage.
A virtual software company platform powered by multiple LLM agents for automated software development.
Market Position
Leader in efficient on-device large models and training toolkits, positioned as an open-source alternative to high-resource LLM providers by focusing on model compression and edge deployment.
Leadership
Founders
Li Dahai
CEO and Co-founder. Previously CTO of Zhihu. Former engineer at Google and Baidu.
Zhiyuan Liu
Co-founder. Associate Professor at Tsinghua University's Department of Computer Science and Technology. Lead researcher at TsinghuaNLP.
Executive Team
Li Dahai
CEO & Co-founder
Former CTO of Zhihu; veteran engineer from Google and Baidu.
Zhiyuan Liu
Co-founder & Principal Scientist
Associate Professor at Tsinghua University; renowned researcher in NLP.
Board of Directors
Founding Story
OpenBMB and its commercial arm ModelBest were born out of the Tsinghua University Natural Language Processing Lab (THUNLP) to bridge the gap between academic research in large models and industrial application, specifically focusing on efficiency and end-side deployment.
Business Model
Revenue Model
Enterprise AI solutions, private model deployments, strategic partnerships (e.g., China Telecom), and licensing on-device AI technology.
Pricing Tiers
Apache 2.0 or custom community licenses for models like MiniCPM.
Private deployment, fine-tuning, and dedicated support.
Target Markets
- Consumer Electronics (Smartphones/IoT)
- Telecommunications
- Enterprise Software
- AI Research Community
- Smartphone AI assistants
- On-device privacy-focused AI
- Efficient large-scale model training
- Automated software engineering (ChatDev)
- Edge computing applications
- China Telecom
- Xiaomi
- Zhihu
- Qualcomm