The best on-demand open-source AI in a sovereign Cloud

Designed for complex tasks that require a large context window and greater precision in logical reasoning.

Architecture optimised for faster inference and reduced energy consumption, freeing up significant computing resources.

Trained on millions of agents and increasingly complex tasks to ensure robust real-world adaptability.

Ideal for multilingual services, public administrations and R&D teams looking for a reliable and adaptable model

Documented data and methods for unprecedented transparency

Compliant with the AI Act and respectful of privacy and intellectual property

A 70B version with performance comparable to today's market leaders

The ideal trade-off between responsiveness and power, designed to excel at logical reasoning, in-depth document analysis and reliable code generation.

Leverages a state-of-the-art architecture to deliver a deep understanding of advanced contexts and complex instructions.

Ideal for advanced conversational agents and business workflows that require great versatility without sacrificing execution speed.

Natively multimodal: turns text, images or mockups into fully functional code.

Built for large-scale development: features an extended context window of up to 256k tokens to handle complex projects

Optimised for vibe coding: a fast, smooth and creative experience designed for developers and product designers

Compatible with agentic workflows: automates analysis, code generation and its end-to-end execution

Optimised for fast and cost-effective deployment, ideal for conversational agents, document analysis and specialised tasks.

Delivers performance comparable to Mistral Small 3.2 24B with minimal resources.

Able to analyse images and provide information based on visual content as well as text.

Advanced multilingual model able to process short queries, paragraphs and long documents of up to 8192 tokens simultaneously

Combines lexical analysis (keywords) and semantic analysis (meaning) for unrivalled ranking accuracy on complex corpora

The ideal solution for enterprise search engines and RAG applications that require a deep understanding of context

Ultra-lightweight architecture (0.6 billion parameters) designed for very low-latency inference and minimal energy consumption

Maintains high relevance accuracy even with a context window extended up to 32768 tokens

Ideal for real-time data streams, autonomous agents and large-scale deployments

The most powerful open-source embedding model on the market

The benchmark for semantic search and retrieval-augmented generation (RAG) tasks

Ideal for advanced use of embedding vectors across various use cases

Exceptional performance, regardless of the text language (100+ languages)

This model is the result of joint work based on a model published by Microsoft

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

Attractive performance for relatively simple tasks, regardless of the text language

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

High energy efficiency to reduce environmental impact

Model trained on over 1 million hours of data

Reduction of transcription errors by up to 20% compared to Whisper V2

Better handling of accents, background noise and complex speech (for example, calls or video conferences)

Improved multilingual support and translation of transcriptions into languages other than English

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

Fast creation of photorealistic images in 1, 2, 4 or 8 steps from a prompt

Works by distillation, which increases energy efficiency while ensuring excellent quality

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

Creation of photos in multiple styles from one or more profile photos

Powerful and flexible: recontextualisation, colourisation, age and gender change, identity mixing...