Mistral-NeMo-Instruct-2407 and Mistral-NeMo-Base-2407 are now available on SageMaker JumpStart
Machine Learning Blog
AWS has announced the availability of Mistral-NeMo-Instruct-2407 and Mistral-NeMo-Base-2407 large language models on Amazon SageMaker JumpStart, offering advanced multilingual AI capabilities.
- 12 billion parameter models developed by Mistral AI and NVIDIA
- Features a 128k token context window with strong reasoning and multilingual performance
- Supports over 10 languages including English, French, Chinese, and Arabic
- Uses Tekken, a new tokenizer offering improved compression efficiency
- Can be deployed easily through SageMaker JumpStart with one-click deployment
The models are available in both base and instruction-tuned versions, demonstrating impressive capabilities in code generation, mathematical reasoning, and language translation tasks.
The AWS News Feed is currently looking for gold sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.
Related articles
2025
2025
2025
2024
The AWS News Feed is currently looking for silver sponsors. If you want to support the AWS community and reach a large audience of AWS professionals, consider sponsoring the AWS News Feed.