The best open source AI on demand in a sovereign cloud

Discover the best open source alternatives to ChatGPT, Gemini, Midjourney or Claude for processing sensitive data in full compliance with European and Swiss law.

LLM

Embedding

Audio

Image

Large language models (LLM)

The best open source alternatives to ChatGPT, Gemini and Microsoft Copilot for interacting, analysing and generating content with AI.

Qwen3-235B-A22B-Instruct-2507

Qwen3-235B-A22B-Instruct-2507

The most powerful

  • Very large-scale model, rivalling GPT-4 or Claude 3 Opus across a broad range of complex tasks

  • Advanced multilingual capabilities

  • Reasoning mode can be enabled to dynamically tailor responses to the context and complexity of queries

Modality

Text to Text

Max. input tokens

262’144

Languages

100+ languages

Function call

Yes

Template category

chat_large

  • Very large-scale model, rivalling GPT-4 or Claude 3 Opus across a broad range of complex tasks

  • Advanced multilingual capabilities

  • Reasoning mode can be enabled to dynamically tailor responses to the context and complexity of queries

Modality

Text to Text

Max. input tokens

262’144

Languages

100+ languages

Function call

Yes

Template category

chat_large

Mistral-Small-3.2-24B-Instruct-2506

Mistral-Small-3.2-24B-Instruct-2506

The most visual

  • Versatile multimodal model, ideal for vision, image analysis, and conversational agents

  • Instant responses with strong contextual understanding

  • Efficient support for all major European languages

Modality

Image-Text to Text

Max. input tokens

128’000

Languages

EN, ES, FR, DE, IT...

Function call

Yes

Template category

vision_medium

  • Versatile multimodal model, ideal for vision, image analysis, and conversational agents

  • Instant responses with strong contextual understanding

  • Efficient support for all major European languages

Modality

Image-Text to Text

Max. input tokens

128’000

Languages

EN, ES, FR, DE, IT...

Function call

Yes

Template category

vision_medium

Gemma-3n-E4B-it

Gemma-3n-E4B-it

The most flexible

  • Small, highly efficient multimodal model that is cost-effective to deploy

  • Optimised for constrained environments and embedded use cases

  • Suitable for applications requiring fast responses in vision or text

Modality

Image-Audio-Text to Text

Max. input tokens

32’000

Languages

140+ languages

Function call

Yes

Template category

omni_small

  • Small, highly efficient multimodal model that is cost-effective to deploy

  • Optimised for constrained environments and embedded use cases

  • Suitable for applications requiring fast responses in vision or text

Modality

Image-Audio-Text to Text

Max. input tokens

32’000

Languages

140+ languages

Function call

Yes

Template category

omni_small

Llama 3.3

Llama 3.3

The most powerful

  • Optimised to handle large amounts of text ensuring consistency across multiple sources

  • Excellent in development, programming and academic research tasks

  • High multilingual flexibility with more than 30 languages supported

  • Suitable for artists and content creation, including storytelling

Modality

Text to Text

Max. input tokens

100’000

Languages

EN, ES, FR, DE, IT...

Function call

Yes

  • Optimised to handle large amounts of text ensuring consistency across multiple sources

  • Excellent in development, programming and academic research tasks

  • High multilingual flexibility with more than 30 languages supported

  • Suitable for artists and content creation, including storytelling

Modality

Text to Text

Max. input tokens

100’000

Languages

EN, ES, FR, DE, IT...

Function call

Yes

Embedding models

The best open-source embedding models to transform your data into intelligent vectors. Improve search accuracy, personalise recommendations, simplify data analysis, explore semantic links and easily classify text.

Bge Multilingual Gemma2

Bge Multilingual Gemma2

The highest quality

  • The most powerful open-source embedding model on the market

  • The benchmark for semantic search and augmented search (RAG) tasks

  • Ideal for advanced use of embedding vectors in a variety of use cases

  • Outstanding performance, whatever language the text is in (100+ languages)

Max. input tokens

8192

Parameters

9.2 B

Dimensions

3584

Languages

EN, ES, FR, DE, IT...

Type

Text

  • The most powerful open-source embedding model on the market

  • The benchmark for semantic search and augmented search (RAG) tasks

  • Ideal for advanced use of embedding vectors in a variety of use cases

  • Outstanding performance, whatever language the text is in (100+ languages)

Max. input tokens

8192

Parameters

9.2 B

Dimensions

3584

Languages

EN, ES, FR, DE, IT...

Type

Text

All MiniLM L12 v2

All MiniLM L12 v2

The best value for money

  • This model is the result of community work based on a model published by Microsoft.

  • Excellent value for money, perfect for prototyping and simple tasks with limited resources

  • Great performance for relatively simple tasks, whatever language the text is in

  • Extreme speed for indexing huge databases or real-time processing

  • High energy efficiency to reduce environmental impact

Max. input tokens

512

Parameters

33 M

Dimensions

384

Languages

EN, ES, FR, DE, IT...

Type

Text

  • This model is the result of community work based on a model published by Microsoft.

  • Excellent value for money, perfect for prototyping and simple tasks with limited resources

  • Great performance for relatively simple tasks, whatever language the text is in

  • Extreme speed for indexing huge databases or real-time processing

  • High energy efficiency to reduce environmental impact

Max. input tokens

512

Parameters

33 M

Dimensions

384

Languages

EN, ES, FR, DE, IT...

Type

Text

Voice recognition

The best open source AI for transcribing audio files into text or generating realistic human voices.

Whisper V3

Whisper V3

For complex transcriptions

  • Model trained on over 1 million hours of data

  • Transcription errors reduced by up to 20% compared with Whisper V2

  • Better handling of accents, background noise and complex speech (e.g., calls or videoconferences)

  • Enhanced multilingual support and translation of transcriptions into languages other than English

Maximum file size

25 MB

Formats supported

mp3, mp4, aac, wav, flac, ogg, opus, wma, m4a

  • Model trained on over 1 million hours of data

  • Transcription errors reduced by up to 20% compared with Whisper V2

  • Better handling of accents, background noise and complex speech (e.g., calls or videoconferences)

  • Enhanced multilingual support and translation of transcriptions into languages other than English

Maximum file size

25 MB

Formats supported

mp3, mp4, aac, wav, flac, ogg, opus, wma, m4a

Image generation and processing

The best open source alternatives to Midjourney, Microsoft Copilot Designer and Gemini for generating, merging or interpreting images.

Photomaker V2

Photomaker V2

Ideal for generating images

  • The best combination of quality and speed in generative AI image creation

  • Fast generation of photo-realistic images in 1, 2, 4 or 8 steps based on prompts

  • Operates by distillation, which increases energy efficiency and ensures excellent quality

  • Optimised for English, with limited knowledge of other languages (FR, DE, ES, IT, etc.)

Max. input tokens

77

Max. image output

5

Languages

EN

Maximum resolution

1024x1024, 1792x1024, 1024x1792

  • The best combination of quality and speed in generative AI image creation

  • Fast generation of photo-realistic images in 1, 2, 4 or 8 steps based on prompts

  • Operates by distillation, which increases energy efficiency and ensures excellent quality

  • Optimised for English, with limited knowledge of other languages (FR, DE, ES, IT, etc.)

Max. input tokens

77

Max. image output

5

Languages

EN

Maximum resolution

1024x1024, 1792x1024, 1024x1792

Flux schnell

Flux schnell

Ideal for modifying and merging portraits of people

  • Create photos in multiple styles from one or more profile photos

  • Powerful and flexible: recontextualisation, colourisation, age and gender change, mix of identities, etc.

Max. input tokens

77

Max. image input

6

Max. image output

5

Languages

EN

Maximum resolution

1024x1024, 1792x1024, 1024x1792

  • Create photos in multiple styles from one or more profile photos

  • Powerful and flexible: recontextualisation, colourisation, age and gender change, mix of identities, etc.

Max. input tokens

77

Max. image input

6

Max. image output

5

Languages

EN

Maximum resolution

1024x1024, 1792x1024, 1024x1792