Beijing Academy of Artificial Intelligence (BAAI)
20 models across 3 families · Latest: BGE Multilingual Gemma2 (2024-06)
Open-source AI fostering global collaboration
Beijing Academy of Artificial Intelligence (BAAI)'s portfolio covers 20 active models across 3 current families, spanning general LLM work. Open a model detail page to compare provider routes and sourced benchmarks.
Covers 0 workload areas across 20 active tracked models; last verified 2026-06-03.
Use it for
- Teams evaluating general LLM work across this lab's releases
- Comparing model families before committing to a flagship
- Migration and pricing follow-ups across 20 tracked models
Do not use it for
- Choosing a hosting provider without opening a model page for price ladders
Active models
20
Current models from this lab, excluding deprecated ones
Active families
3
Current model families from this lab
Open catalog
8 open
8 open source / 0 open weights
Lowest output price
Not tracked
No provider output pricing linked yet
Latest dated release
2024-06-29
BGE Multilingual Gemma2
Freshness
2026-06-03
Researched 45d ago
Information
Release cadence
Showing 5 recent dated releases (full timeline below). Latest: BGE Multilingual Gemma2 (2024-06-29).
Where this lab wins
Not enough capability or benchmark coverage yet to call strengths for this lab.
Flagship quality / price signal
Flagship: Aquila Chat 2 70B Expressive (best sourced coding quality-per-dollar in this portfolio).
Quality-per-dollar unavailable for this flagship — benchmark coverage or output token pricing is still missing.
Beijing Academy of Artificial Intelligence (BAAI) is a Chinese AI research organization founded in 2018. Open-source AI fostering global collaboration. Beijing Academy of Artificial Intelligence (BAAI) ships 3 model families totaling 20 models, with the most recent release BGE Multilingual Gemma2 in 2024-06. Notable families include BGE, Aquila 2, and Aquila. Use it as a stable reference for lab background, release coverage. View official API endpoints, benchmark performance, and coding/agent fit for every Beijing Academy of Artificial Intelligence (BAAI) model.
About
The Beijing Academy of Artificial Intelligence (BAAI) is a prominent non-profit research organization established in 2018 in Beijing, China. Founded through a collaborative effort among leading AI companies, universities, and research institutes within China, BAAI is dedicated to advancing AI research and development, emphasizing technologies that benefit people and the planet. The organization counts among its founding members such influential entities as Baidu, ByteDance, Megvii, Meituan, and Xiaomi, along with esteemed universities like Tsinghua University and Peking University. BAAI is known for its notable contributions to the field of generative AI and Large Language Models (LLMs), particularly with the development of the Wu Dao model series. Wu Dao 2.0, an expansive multimodal generative AI model with 1.75 trillion parameters, marked a significant milestone in AI capabilities. However, due to challenges like high costs and regulatory constraints, subsequent models such as Wu Dao 3.0 have been developed on a smaller scale, aiming to enhance accessibility for Chinese startups and SMEs. These include the Wu Dao Aquila series, with models like AquilaChat for dialogue generation and AquilaCode for text-to-code applications, all designed to support bilingual capabilities. Beyond model creation, BAAI's role extends deeply into the wider AI ecosystem. They launched FlagOpen, a comprehensive open-source platform designed to support a variety of technological frameworks and heterogeneous chips to foster innovation in large models. FlagOpen 2.0, a significant upgrade, targets comprehensive improvements in modeling, algorithms, and system architecture, aspiring to set a foundational standard for developing large-scale AI models. Moreover, BAAI underscores its commitment to open-source collaboration through the release of extensive datasets like InfinityInstruct and IndustryCorpus. These resources address data availability issues and enhance the robustness of AI models. Their initiatives in creating research centers focusing on health, biocomputing, life simulation, and AI ethics highlight their dedication to responsible AI, aiming to confront societal challenges and foster sustainable development. BAAI’s persistent push towards open-source methodologies and data sharing solidifies its status as a leader in AI innovation. By promoting collaboration within the global AI community, BAAI accelerates progress and propels exploratory ventures in generative AI and LLMs, ensuring they are leveraged responsibly to address critical global issues.
Featured models
| Model | Released | Context | Input price ($/1M) | Output price ($/1M) | License | Openness |
|---|---|---|---|---|---|---|
| BGE Multilingual Gemma2 | 2024-06-29 | 4k | - | - | MIT | Open source |
| BGE Reranker V2 M3 | 2024-03-18 | 8k | $0.01 | - | MIT | Open source |
| BGE M3 | 2024-01-27 | 8k | $0.01 | - | MIT | Open source |
Model families
Recent releases
- BGE Multilingual Gemma2- 2024-06-29
- BGE Reranker V2 M3- 2024-03-18
- BGE M3- 2024-01-27
- Aquila Chat 2 70B Expressive- 2023-11-02
- Aquila Chat 2 34B- 2023-11-02
Top comparisons
- Aquila 2 7B vs Bielik 11B v2.6 Instruct37
- Aquila 2 34B vs Hunyuan Hy3 Preview12
- Aquila 2 7B vs Falcon 180B9
- Aquila 2 7B vs Xiaomi MiMo-V2.59
- Aquila 2 7B vs Dracarys Llama 3.1 70B Instruct9
- Aquila Chat 2 70B Expressive vs Dracarys Llama 3.1 70B Instruct8
- Aquila 2 7B vs Llama 3.1 Swallow 8B Instruct8
- Aquila 2 34B vs Llama 2 7B Chat7
FAQ
Who founded Beijing Academy of Artificial Intelligence (BAAI) and when?
Beijing Academy of Artificial Intelligence (BAAI) was founded in 2018 and is associated with Beijing, China.
What models has Beijing Academy of Artificial Intelligence (BAAI) released?
Beijing Academy of Artificial Intelligence (BAAI) ships 20 models across 3 families: BGE, Aquila 2, and Aquila.
Is Beijing Academy of Artificial Intelligence (BAAI)'s technology open source?
Some Beijing Academy of Artificial Intelligence (BAAI) models are open-weight (BGE Multilingual Gemma2, BGE Reranker V2 M3, and BGE M3); others are proprietary (Aquila Chat 2 70B Expressive, Aquila Chat 2 34B, and Aquila Chat 2 7B).
Where is Beijing Academy of Artificial Intelligence (BAAI) headquartered?
Beijing Academy of Artificial Intelligence (BAAI) is headquartered in Beijing, China.
What is Beijing Academy of Artificial Intelligence (BAAI) known for?
Open-source AI fostering global collaboration. Its most prominent tracked family is BGE.
How can I access Beijing Academy of Artificial Intelligence (BAAI)'s models?
Beijing Academy of Artificial Intelligence (BAAI)'s models are available via Cloudflare Workers AI and Novita AI.
Explore related pages
Last reviewed: 2026-06-03. Data sourced from public lab announcements and provider documentation.