Mistral AI’s latest offering, Mistral Large 2 (24.07), is now generally available on Amazon Bedrock. This new foundation model (FM) represents a significant leap forward in AI capabilities, delivering enhanced performance across multilingual tasks, mathematics, reasoning, and coding. In this post, we’ll explore the key benefits and capabilities of Mistral Large 2, supported by practical examples.
Overview of Mistral Large 2 Mistral Large 2 is an advanced large language model (LLM) designed with state-of-the-art reasoning, knowledge, and coding capabilities. It supports dozens of languages, including English, French, German, Spanish, Italian, Chinese, Japanese, Korean, Portuguese, Dutch, Polish, Arabic, and Hindi. A major focus during its development was to reduce the model’s tendency to hallucinate or generate plausible-sounding but factually incorrect information. This was achieved through fine-tuning, making the model more cautious and accurate in its responses. Mistral Large 2 is also capable of acknowledging when it lacks sufficient information to provide a confident answer.
The model is proficient in coding, having been trained on over 80 programming languages, such as Python, Java, C, C++, JavaScript, Bash, Swift, and Fortran. It can natively call functions and output JSON, facilitating seamless interaction with external systems, APIs, and tools. Additionally, Mistral Large 2 boasts advanced reasoning and mathematical capabilities, making it a powerful tool for complex logical and computational tasks.
Mistral Large 2 supports a context window of 128,000 tokens, significantly larger than its predecessor. This extended context window allows the model to process and understand longer texts, such as entire documents or code files, without losing coherence. The model is currently available in the us-west-2 AWS Region.
Getting Started with Mistral Large 2 on Amazon Bedrock To access Mistral Large 2, request model access through the Amazon Bedrock console. For detailed instructions, refer to the Amazon Bedrock documentation.
To test the model, navigate to the Text or Chat sections under Playgrounds in the Amazon Bedrock console. Select Mistral as the category and Mistral Large 24.07 as the model. You can also access the model programmatically using the AWS Command Line Interface (AWS CLI) and AWS SDKs. Here’s an example:
aws bedrock-runtime invoke-model \
--model-id mistral.mistral-large-2407-v1:0 \
--body "{\"prompt\":\"<s>[INST] this is where you place your input text [/INST]\", \"max_tokens\":200, \"temperature\":0.5, \"top_p\":0.9, \"top_k\":50}" \
--cli-binary-format raw-in-base64-out \
--region us-west-2 \
invoke-model-output.txt
Capabilities of Mistral Large 2
Increased Context Window Mistral Large 2’s 128,000-token context window allows it to handle longer texts more effectively. This is crucial for tasks like code generation, documentation analysis, and other applications requiring extensive text processing.
Generating JSON and Tool Use The model now offers native JSON output, simplifying integration with various applications. For example, using the Converse API, you can generate structured JSON responses.
Multilingual Support Mistral Large 2 supports many languages, including character-based languages like Chinese, Japanese, and Korean. This allows for creating applications catering to diverse linguistic audiences. For example, translating customer emails into different languages.
Coding Tasks Mistral Large 2 excels in coding tasks across various languages. Here’s an example of generating a Python function to find palindrome prime numbers.
General Performance and Licensing Mistral Large 2, with 123 billion parameters, sets new benchmarks for performance and cost-efficiency. It is available under the Mistral Research License for research and non-commercial use. For commercial use, a Mistral Commercial License is required.
Conclusion Mistral Large 2 on Amazon Bedrock offers enhanced capabilities in multilingual support, coding, reasoning, and more. It sets new standards for performance and accuracy, making it an invaluable tool for developers and businesses alike. To get started, visit the Amazon Bedrock console and explore the potential of Mistral Large 2 today.
If you want to run the best open source model Llama 3.1 405B for your enterprise via Amazon Bedrock, click here.
For more information, visit the Amazon Bedrock documentation or contact us at hello@deskinvestor.com to get your enterprise AI ready.