Mistral-on-AWS

A collection of notebooks and samples to get started with Mistral models on AWS. Open a PR if you would like to contribute! 🔀

What's New ⭐⭐

New Model Launches on Amazon Bedrock

Mistral Large 3 (675B)

Mistral's flagship frontier model is now available on Amazon Bedrock! A sparse Mixture-of-Experts architecture with 41B active parameters out of 675B total, delivering frontier-level multimodal performance under the Apache 2.0 license.

• Mistral Large 3 Capabilities Guide - Text generation, multilingual (40+ languages), vision, and complex reasoning

Specification	Details
Model ID	mistral.mistral-large-3-675b-instruct
Architecture	Sparse MoE (41B active / 675B total)
Multimodal	Yes (Text + Vision)
License	Apache 2.0

Ministral Models (3B, 8B, 14B)

Compact yet powerful models optimized for edge deployment and cost-efficient inference, now available on Amazon Bedrock!

• Ministral Capabilities Guide - Model comparison, text generation, reasoning, code, JSON output, and vision

Model	Model ID	Best For
Ministral 3B	mistral.ministral-3-3b-instruct	Edge devices, ultra-low latency
Ministral 8B	mistral.ministral-3-8b-instruct	Balanced performance/efficiency
Ministral 14B	mistral.ministral-3-14b-instruct	Complex reasoning, code generation

Building AI Agents with Strands

Learn how to build production-ready AI agent systems using the Strands Agents SDK with Mistral models on Bedrock.

• Strands Agents with Mistral - Multi-agent orchestration, streaming, hooks, session persistence, and expert panels

---

Model Capabilities & Use Cases

Pixtral

• Comprehensive Capabilities Guide
• Deployment Options:
  • SageMaker Real-time Inference
  • Bedrock Marketplace Integration

Mistral Models

• Large 3: Frontier Multimodal Model Guide - 675B MoE with vision
• Small 3: Model Overview & Capabilities
• NeMo: Comparative Analysis & Benchmarks
• Ministral: 3B, 8B, 14B Models Guide - Compact models for edge and cost-efficient deployment

💡 Note: All notebooks include detailed explanations, code samples, and best practices for implementation.

Getting Started 🔌

1. Please visit Amazon Bedrock user guide on how to enable model access.
2. The notebooks are executed from SageMaker studio with Data Science 3.0 image.

Security

See CONTRIBUTING for more information.

Distributors

• AWS
• Mistral

License

This library is licensed under the MIT-0 License. See the LICENSE file.