SAIL Models
Sovereign AI Models Directory
A small, curated set of sovereign AI models, tagged by the interest each one primarily serves. The point is not to rank models but to show that βsovereignβ means different things for each: a regional-language model, an open public good, and a national champion are answers to different questions.
Featured Models
GPT-SW3 πΈπͺ Sweden
GPT-SW3 is a large language model trained on Swedish text data, developed by AI Sweden in collaboration with RISE and the Wallenberg AI, Autonomous Systems and Software Program (WASP). It represents one of the first major sovereign AI initiatives in the Nordic region.
Serves which interest?
Cultural identity & values
Middle-power alliance
40B
Parameters
Swedish
Primary Language
Open
License
Decoder
Architecture
SEA-LION πΈπ¬ Singapore
SEA-LION (Southeast Asian Languages in One Network) is a family of large language models designed specifically for Southeast Asian languages. Developed by AI Singapore, it supports 11 languages including Malay, Indonesian, Thai, Vietnamese, and Tagalog.
Serves which interest?
Middle-power alliance
Cultural identity & values
7B-128B
Parameters
11 Languages
Multilingual
Apache 2.0
License
Decoder
Architecture
Apertus π¨π Switzerland
Apertus is a fully open, transparent, multilingual language model developed by ETH Zurich. It represents a significant advancement in open-source sovereign AI, with complete transparency in training data, methodology, and deployment.
Serves which interest?
Cultural identity & values
Middle-power alliance
70B
Parameters
Multilingual
Languages
Fully Open
License
Decoder
Architecture
Mistral AI Models π«π· France
Mistral AI has developed a series of high-performance language models including Mistral 7B, Mixtral 8x7B, and Mistral Large. While a private company, Mistral AI represents significant French sovereign AI capabilities and has received substantial public support.
Serves which interest?
Industrial policy
Cultural identity & values
7B-70B
Parameters
Multilingual
Languages
Mixed
License
Decoder
Architecture
AuroraGPT πΊπΈ United States
AuroraGPT is a science-focused language model developed by Argonne National Laboratory. It's designed specifically for scientific applications and represents a significant public sector AI initiative in the United States.
Serves which interest?
Industrial policy
Enterprise & procurement
Science
Domain Focus
English
Primary Language
Open
License
Research
Use Case
Contributing
This is an illustrative selection, not an exhaustive index. Know of a sovereign AI model that sharpens the picture - especially one that serves an interest not yet represented here? Submit information or open a pull request to add it.