French AI startup Mistral has introduced Mistral Code, its new AI-powered coding assistant tailored for large enterprises. The announcement, made on Wednesday, marks Mistral’s strongest move yet into the competitive AI coding tools space.
According to Mistral’s official newsroom post, “Mistral Code is an AI-powered coding assistant that bundles powerful models, an in-IDE assistant, local deployment options, and enterprise tooling into one fully supported package.”
Mistral Code is based on the open-source project Continue but adds enterprise-grade features such as role-based access control (RBAC), audit logs, fine-tuning capabilities, and usage analytics.
“Our goal with Mistral Code is simple: deliver best-in-class coding models to enterprise developers, enabling everything from instant completions to multi-step refactoring—through an integrated platform deployable in the cloud, on reserved capacity, or air-gapped on-prem GPUs,” Mistral wrote.
Mistral Code addresses common concerns about AI coding tools
Many companies have hesitated to adopt AI coding tools due to security risks, lack of customization, and compliance headaches. Mistral states that it built the platform following extensive interviews with VPs of engineering, platform leads, and CISOs. The same four blockers kept coming up, according to the company:
- Limited connectivity to internal repos and services.
- No ability to customize models.
- Shallow capabilities ending at autocomplete.
- Fragmented vendor relationships and SLAs.
Mistral Code aims to address these issues by providing a single, integrated solution that keeps everything, from AI models to actual code, within a company’s systems.
How Mistral Code works
The assistant is powered by four of Mistral’s own AI models, including Codestral, Codestral Embed, Devstral, and Mistral Medium. Developers can fine-tune these models on their private codebases, something that isn’t possible with closed systems like GitHub Copilot.
Mistral Code also supports more than 80 programming languages and can work with files, Git changes, terminal outputs, and issue trackers. For IT teams, Mistral includes an admin dashboard with controls for access, logging, and monitoring usage.
The race for enterprise AI coding
Mistral Code enters a crowded market dominated by tools such as Anysphere’s Cursor, GitHub Copilot, OpenAI Codex, Amazon’s CodeWhisperer, and other players in the AI development space. However, Mistral’s key differentiator is its focus on enterprise security and compliance, an area where many rivals struggle.
Still, the company faces stiff competition in a crowded space. A recent Stack Overflow survey found that “76% of devs used or were planning to use AI tools in their development processes last year,” making the market ripe for innovation.
How to try Mistral Code
Mistral Code is now in private beta for JetBrains IDEs and Microsoft’s Visual Studio Code, with general availability expected soon. To try Mistral Code, enterprises can request access from their Mistral account teams. Mistral says it offers three deployment options: serverless, cloud, or self-hosted on-premises GPUs.