Mistral Small 4 is Mistral AI’s new __hybrid model__, merging Magistral’s __reasoning__ capabilities, Pixtral for __multimodal__ and Devstral for __code agentic__. __Mixture of Experts__ architecture with 128 experts of which 4 active per token, 119 billion total parameters for 6 billion active. Distributed under __Apache 2.0 license__, it integrates via Mistral API, La Plateforme and Le Chat, and powers both conversational assistants and autonomous developer __agents__.
What is Mistral Small 4?
Mistral Small 4 is an open source hybrid language model designed by Mistral AI. Its architecture is based on a Mixture of Experts with 128 experts of which 4 are activated per token, for a total of 119 billion parameters of which only 6 billion are active at inference. This approach brings significant energy and economic efficiency without sacrificing model depth. The model natively accepts text and image inputs, supports configurable reasoning effort and excels particularly on agentic and code tasks. The Apache 2.0 license authorizes its download, modification and deployment, including for commercial uses.
Main Features
Mistral Small 4 brings together several technical advances. Its Mixture of Experts architecture allows scaling the number of total parameters without penalizing inference cost, making the model more accessible to deploy on reasonably-sized infrastructures. Native multimodal capability removes the need for an additional vision model to jointly process text and images. Configurable reasoning effort allows arbitrating between speed and depth of analysis depending on the use case. Specialization on code agentic tasks, inherited from Devstral, makes it an excellent engine for developer copilots and autonomous agents. The model is available via Mistral’s Le Chat, La Plateforme and official API, but also downloadable for on-premise deployments. Mistral AI joined the NVIDIA Nemotron Coalition as a founding member, securing the GPU optimization ecosystem around the model.
Use Cases
Mistral Small 4 use cases span a wide spectrum. Developers integrate it into code copilots and autonomous agents capable of executing complex technical tasks. Data science teams use it for structured reasoning, document analysis and information extraction on large corpora. Customer service deploys conversational assistants capable of understanding screenshots or scanned documents thanks to multimodality. Sovereign organizations, notably in the public, financial or health sectors, adopt it to benefit from AI deployable on-premise under an open license. Startups build their products on the Mistral API layer to combine controlled cost and model quality. Researchers appreciate the accessibility of the model for evaluating new alignment methods, fine-tuning and quantization.
Advantages
The main benefit of Mistral Small 4 lies in the convergence of three model families into one. This unification simplifies the architecture of AI applications by removing the need for routing between multiple specialized models. The Apache 2.0 license brings major strategic independence for organizations not wishing to depend on proprietary APIs. The efficiency of the MoE architecture enables consideration of controlled-cost deployments while benefiting from high quality. The maturity of the Mistral ecosystem, with its Plateforme, API and Le Chat, offers multiple entry points depending on teams‘ technical level. Finally, the model’s European origin is a decisive argument for stakeholders sensitive to GDPR compliance and digital sovereignty.
Pricing
Mistral Small 4 is made available free under Apache 2.0 license, which authorizes on-premise deployment without licensing fees. Access via the Mistral API is billed on usage according to consumed tokens at competitive rates aligned with the model’s efficiency strategy. Mistral’s Le Chat offers free access with quotas to test model capabilities, as well as paid plans for intensive use. La Plateforme integrates monitoring and fine-tuning tools for professional teams. Overall, Mistral Small 4 offers excellent pricing flexibility, from free on-premise deployment to serverless use via API, through enterprise plans for high-volume organizations.
Conclusion
Mistral Small 4 confirms Mistral AI’s strategy: offering open, performant and technically differentiated models. Its multimodal versatility and configurable reasoning effort make it a relevant choice for many use cases, from developer copilot to multimodal assistant. The Apache 2.0 license and on-premise deployment possibility will particularly appeal to European stakeholders and sovereign organizations. For technical teams seeking a flexible, performant model independent of American giants, Mistral Small 4 is today a major reference in the open source AI landscape.