Mistral AI have announced Mistral Code, an AI-powered coding assistant designed for enterprise software engineering teams requiring secure, compliant AI coding solutions. Built on the open-source Continue project, Mistral Code bundles powerful models, in-IDE assistance, local deployment options, and enterprise tooling into a fully supported package aimed at enabling developers to "10X their productivity with the full backing of their IT and security teams."
The platform addresses four recurring enterprise blockers identified through discussions with VPs of engineering, platform leads, and CISOs: limited connectivity to proprietary repositories and internal services, minimal customisation of underlying models or prompts, shallow task coverage ending at autocomplete instead of multi-step work, and fragmented SLAs across multiple vendors for plugins, models, and infrastructure.
Mistral Code operates through four state-of-the-art coding models: Codestral for fill-in-the-middle code autocomplete, Codestral Embed for code search and retrieval, Devstral for agentic coding, and Mistral Medium for chat assistance. The system supports 80+ programming languages and can reason over files, Git diffs, terminal output, and issues. Unlike typical SaaS copilots, all components reside within the customer's enterprise boundary with single-vendor SLA coverage.
Customers can fine-tune or post-train underlying models on private repositories or distill lightweight variants—capabilities unavailable in closed copilots tied to proprietary APIs. The platform offers cloud, reserved capacity, or air-gapped on-premises GPU deployment options, with rich admin console controls including granular platform controls, deep observability, seat management, and usage analytics.
Enterprise adoption includes Abanca, a leading Spanish and Portuguese bank, deploying Mistral Code at scale in hybrid configuration allowing cloud prototyping, while keeping core banking code on-premises. SNCF, France's national railway company, empowers 4,000 developers through Mistral Code Serverless. Capgemini serves as the first global systems-integrator partner, deploying on-premises solutions for 1,500+ developers in regulated industry client projects.
Private beta have been launched for JetBrains IDEs and VSCode with general availability planned soon. The company acknowledges Continue community contributions while adding enterprise-grade features including fine-grained RBAC, audit logging, and issue resolution metrics.
Mistral Code enables enterprises to maintain code security within their boundaries while accessing frontier AI capabilities. The platform addresses regulatory compliance requirements through on-premises deployment while providing comprehensive administrative controls and observability for IT teams managing developer productivity initiatives.
Mistral AI positions itself as the enterprise alternative to mainstream coding copilots by offering deployment flexibility and model customisation unavailable from proprietary solutions. The single-vendor approach simplifies procurement and support while addressing security concerns that prevent enterprise adoption of cloud-based coding assistants. Success with major enterprises like SNCF and Abanca demonstrates viable market demand for secure, customisable AI coding solutions.