Mistral AI: New Coding Models for Enhanced Development

Mistral Launches Devstral 2: A New AI Model for Coding

Mistral, a French AI startup, has unveiled Devstral 2, its latest AI model specifically engineered for coding tasks. This launch signifies the company's ambition to compete with larger AI organizations like Anthropic and other leading Large Language Models (LLMs) focused on code generation.

Following the Mistral 3 Family

The introduction of Devstral 2 comes shortly after the release of the Mistral 3 family of open-weight models. This demonstrates Mistral’s commitment to narrowing the gap with its more established and financially robust AI competitors.

Introducing Mistral Vibe: Natural Language Code Automation

Mistral is also entering the realm of ‘vibe-coding,’ a trend popularized by companies such as Cursor and Supabase. They are doing so with Mistral Vibe, a new command-line interface (CLI) designed to streamline code automation through natural language processing.

Mistral Vibe provides tools for file management, code searching, version control, and executing commands.

Context Awareness for Business Applications

Mistral AI is prioritizing the value of context awareness, which is especially crucial for practical business applications. Like its AI assistant, Le Chat, which retains memory of past user interactions to refine its responses, Vibe CLI offers persistent history.

Furthermore, it can analyze file structures and Git statuses to establish context and improve its performance.

Devstral 2: Specifications and Deployment

This emphasis on production-ready workflows is reflected in Devstral 2’s substantial resource requirements. Deploying the model necessitates a minimum of four H100 GPUs or equivalent hardware, and it comprises 123 billion parameters.

A smaller version, Devstral Small, is also available, containing 24 billion parameters, making it suitable for local deployment on standard consumer-grade computers.

Licensing and Pricing Details

The two models differ in their open-source licensing terms. Devstral 2 is released under a modified MIT license, while Devstral Small utilizes the Apache 2.0 license.

Pricing structures also vary. Currently, Devstral 2 is accessible via the company’s API at no cost. Following the initial free period, API usage will be priced at $0.40/$2.00 per million tokens (input/output) for Devstral 2, and $0.10/$0.30 for Devstral Small.

Partnerships and Availability

Mistral has collaborated with agent tools Kilo Code and Cline to make Devstral 2 available to users. Mistral Vibe CLI is offered as an extension within Zed, enabling its use directly within the IDE.

Company Valuation and Funding

As Europe’s leading AI lab, Mistral is currently valued at €11.7 billion (approximately $13.8 billion). This valuation follows a Series C funding round spearheaded by Dutch semiconductor firm ASML, which invested €1.3 billion (approximately $1.5 billion) in September.