Mistral Docs - Mistral AI

2026-04-17 · Source: mistral.ai via Google News · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Software Development & Engineering, Robotics & Autonomous Systems · Depth: Intermediate, medium

Summary

Mistral AI provides an extensive overview of its available models, categorized into Featured, Frontier (Generalist and Specialist), Other, and Legacy/Deprecated. Featured models include Mistral Large 3, Devstral 2, and Mistral Medium 3.5, with the latter two optimized for agentic and coding tasks. Frontier Generalist models like Mistral Large 3 Open and Ministral 3 14B Open offer broad capabilities, while Specialist models such as Voxtral TTS Open and Codestral Embed Premier are fine-tuned for specific domains like text-to-speech, code agents, or OCR. The list also details various other supported models and a comprehensive table of legacy models, specifying their versions, API names, deprecation/retirement dates, and recommended alternatives like Mistral Small 4 or Devstral 2.

Key takeaway

For AI Architects evaluating model portfolios, review Mistral AI's comprehensive model overview to align specific project needs with the appropriate generalist or specialist models. Pay close attention to the deprecation schedule for legacy models to plan timely migrations and ensure your applications remain supported and performant with recommended alternatives like Mistral Large 3 or Devstral 2.

Key insights

Mistral AI offers a diverse portfolio of generalist and specialist models, including multimodal and code-focused agents.

Principles

• Models are categorized by capability and lifecycle status.
• Specialized models target specific domain tasks.
• Legacy models have clear deprecation paths and alternatives.

In practice

• Identify suitable models based on task requirements (generalist vs. specialist).
• Consult deprecation schedule for migrating off legacy models.
• Explore multimodal models for combined text and vision tasks.

Topics

• Mistral AI Model Portfolio
• Multimodal Models
• Code Generation Agents
• Text-to-Speech Synthesis
• Optical Character Recognition

Best for: CTO, VP of Engineering/Data, AI Architect, AI Engineer, Machine Learning Engineer, Director of AI/ML

Related on AIssential

• Mistral Medium 3.5 - Mistral AI — Mistral Medium 3.5 is a multimodal, open-weight model optimized for agents and coding with extensive features.
• Mistral Next - Mistral AI — Mistral Next is a new prototype model designed for concise outputs, replacing Mistral Large 3.
• [AINews] The End of Finetuning — Finetuning is declining for general AI engineering but remains critical for top-tier applications and open models.
• LWiAI Podcast #234 - Opus 4.6, GPT-5.3-Codex, Seedance 2.0, GLM-5 — The AI sector is rapidly advancing with new models, generative media tools, and significant venture capital investments.
• Introducing Mistral Small 4 — Mistral Small 4 unifies multiple specialized AI capabilities into a single, versatile model.
• Data Machina #262 — New models and frameworks are advancing LLM efficiency, reasoning, and long-form content generation.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by mistral.ai via Google News.