Introducing Mistral.mistral-large-2407-v1:0 on Amazon Bedrock
The Mistral AI Large 2 (24.07) model, known as mistral.mistral-large-2407-v1:0, is a powerful new addition to the Amazon Bedrock platform. This model is available in the us-west-2 AWS Region and can be deployed as a fully managed endpoint on AWS Bedrock, a serverless service.
Capabilities and Improvements
- Multilingual Support: The model supports dozens of languages, including English, French, German, Spanish, Italian, Chinese, Japanese, Korean, Portuguese, Dutch, Polish, Arabic, and Hindi.
- Reasoning and Knowledge: Enhanced reasoning capabilities and training to minimize hallucinations provide more reliable and accurate outputs.
- Coding Capabilities: Proficient in over 80 programming languages, such as Python, Java, C, C++, JavaScript, Bash, Swift, and Fortran.
- Agentic Capabilities: Can natively call functions and output JSON, enabling seamless interaction with external systems, APIs, and tools.
Technical Specifications
- Context Window: Supports a context window of 128,000 tokens, significantly larger than the previous version (32,000 tokens).
- JSON Output: Offers a native JSON output mode, making it easier to integrate responses into applications.
Getting Started
To use the model, you need an AWS account, an IAM principal with sufficient permissions, and a local code environment set up with the AWS CLI and boto3 library. You must request access to the model through the Amazon Bedrock console and configure your authentication credentials properly.
Usage Example
The model can be accessed using the AWS Bedrock Converse API. Below is an example of how to invoke the model using Python:
import boto3
import json
bedrock = boto3.client('bedrock-runtime', 'us-west-2')
response = bedrock.converse(
modelId='mistral.mistral-large-2407-v1:0',
messages=[{'role': 'user', 'content': [{'text': 'Which LLM are you?'}]}]
)
print(json.dumps(json.loads(response['body']), indent=4))
