Blog Post

Microsoft Foundry Blog
2 MIN READ

Introducing Kimi K3 through Fireworks AI on Microsoft Foundry

davevoutila's avatar
davevoutila
Icon for Microsoft rankMicrosoft
Jul 28, 2026

Need to process entire codebases, lengthy legal documents, or hours of meeting transcripts in a single prompt? Kimi K3's 1 million-token context window—one of the largest available—is now deployable on Microsoft Foundry through Fireworks AI. Here's what that massive context means for your applications.

What makes Kimi K3 Different

Kimi K3 is an open-weight model from Moonshot AI featuring 2.8 trillion parameters—making it one of the largest open models available. Its 1 million-token context window means you can process approximately 750,000 words (roughly 10 full novels or an entire codebase) in a single request. Fireworks AI provides the optimized inference infrastructure, while Microsoft Foundry handles enterprise deployment and governance.

Use Cases

  • Full repository code analysis: Analyze entire codebases for security vulnerabilities or refactoring opportunities. For example, you could pass your entire Node.js application (package.json, all source files, and dependencies) in a single prompt and ask Kimi K3 to identify unused imports, potential SQL injection points, or suggest architectural improvements—no chunking or retrieval augmentation required.'
  • Long-document summarization: Process legal contracts, research papers, or compliance documents without chunking
  • Extended conversation memory: Build chatbots that remember entire conversation histories

Pricing and Deployment

 

Model Name

Deployment Type

Input/1M tokens

Output/1M tokens

Cached Input/1M tokens

FW Kimi K3

 Data Zone

$3.30

$16.50

$0.33

 

Customers can deploy Kimi K3 in Microsoft Foundry via Fireworks, subject to applicable availability, deployment, and pricing terms.

 

Get Started Today

 

Updated Jul 28, 2026
Version 1.0

1 Comment

  • Evilz's avatar
    Evilz
    Copper Contributor

    Do you know if it will become availlable as "Direct from Azure models" on all region (francecentral for me) ?
    Thank