Feature Drop: Mistral Small 4

Update (April 28, 2026): Mistral AI's newer flagship model, Mistral Medium 3.5, is now available in Omnifact for complex reasoning and agentic coding.

Mistral Small 4 is now available on Omnifact. The latest release in Mistral AI's Small family unifies instruct, coding, and conversational capabilities into a single efficient model, giving teams high-speed performance at a fraction of the cost of flagship tiers.

Building on our original Mistral AI integration, Small 4 expands Omnifact's European provider lineup. It serves as an economical default for high-volume tasks, routine analysis, and fast chat without sacrificing response quality.

Key Capabilities of Mistral Small 4

Mistral Small 4 delivers significant architectural upgrades over earlier Small generations. Core strengths include:

• High-Efficiency Performance: Tuned for low latency and high throughput, Small 4 delivers fast responses for everyday questions, drafting, and conversational assistance.
• Expanded 256k Context Window: Read and synthesize lengthy documents, meeting transcripts, and project notes in a single prompt.
• Multimodal Input Support: Process images, diagrams, and document screenshots alongside text for quick visual Q&A.
• Cost-Effective Processing: At a fraction of the token cost of frontier models, Small 4 makes large-scale classification, extraction, and batch jobs economical.
• European Data Hosting: Operated by Mistral AI in the EU, ensuring your data remains within European borders under strict GDPR compliance.

When to Choose Mistral Small 4

Mistral Small 4 is the ideal choice for everyday, cost-sensitive work across your organization:

• Drafting emails, summaries, and meeting action items.
• Categorizing incoming tickets, customer inquiries, and feedback.
• Extracting structured information from reports and uploaded documents in Spaces.
• Providing a responsive, daily assistant for team members with low latency requirements.

When your projects require deep architectural design, complex code refactoring, or multi-step agentic execution, Mistral Medium 3.5 provides the advanced reasoning depth needed for those tasks.

Get Started

Mistral Small 4 is available immediately across all Omnifact plans.

1. Open your Team Settings if you are a team admin.
2. Navigate to the Models tab.
3. Turn on Mistral Small 4 to make it accessible to your team.

Once enabled, members can choose Mistral Small 4 from the model selector in chat or set it as the default model in their Spaces.