Ministral-3-3B-Instruct-2512 and Ministral-3-8B-Instruct-2512 models now available on Amazon SageMaker JumpStart

New Models Available on Amazon SageMaker JumpStart

Mistral AI's Ministral-3-3B-Instruct-2512 and Ministral-3-8B-Instruct-2512 models are now available on Amazon SageMaker JumpStart, expanding the portfolio of foundation models available to AWS customers. These models are designed for edge deployment and resource-constrained environments, offering high-performance, scalable AI solutions.

Ministral-3-3B-Instruct-2512

• Features: Ultra-lightweight edge deployment with multimodal understanding
• Components: 3.4B language model and 0.4B vision encoder
• Memory: Fits in 8GB of VRAM in FP8
• Context Window: 256K-token
• Capabilities: Vision analysis, multilingual instruction following, system-prompt adherence, native function calling with JSON output
• License: Apache 2.0

Ministral-3-8B-Instruct-2512

• Features: Frontier-class capabilities in a compact 8B form factor
• Components: 8.4B language model and 0.4B vision encoder
• Memory: Fits in 12GB of VRAM in FP8
• Attention Pattern: Interleaved sliding-window for faster, memory-efficient inference
• Capabilities: Vision, multilingual, agentic, and function-calling capabilities with stronger reasoning and generation performance

What to do

• Navigate to the SageMaker JumpStart model catalog in the SageMaker console
• Use the SageMaker Python SDK to deploy the models to your AWS account

For more information, see the Amazon SageMaker JumpStart documentation.

If you need further guidance on AWS, our experts are available at AWS@westloop.io. You may also reach us by submitting the Contact Us form.