What’s new in Mistral Small 3.1
Mistral Small 3.1 is the best model in its weight class. Building upon the foundation of Mistral Small 3, this new model boasts improved text performance, enhanced multimodal understanding, and an expanded context window of up to 128k tokens. Mistral Small 3.1 outperforms comparable models like Gemma 3 and GPT-4o Mini while delivering inference speeds of 150 tokens per second. The model is released under an Apache 2.0 license, making it open source.
Modern AI applications require a blend of capabilities including text handling, multimodal input understanding, support for multiple languages, and management of long contexts, all with low latency and cost efficiency. Mistral Small 3.1 is described as the first open-source model that surpasses the performance of leading small proprietary models across all these dimensions.
Mistral Small 3.1 is designed to handle a wide range of generative AI tasks, including instruction following, conversational assistance, image understanding, and function calling. It provides a solid foundation for both enterprise and consumer-grade AI applications.
Key features and capabilities of the model:
- Lightweight: Can run on a single RTX 4090 or a Mac with 32GB RAM, making it suitable for on-device use cases.
- Fast-response conversational assistance: Ideal for virtual assistants and other applications where quick, accurate responses are essential.
- Low-latency function calling: Capable of rapid function execution within automated or agentic workflows.
- Fine-tuning for specialized domains: Can be fine-tuned to specialize in specific domains, such as legal advice, medical diagnostics, and technical support.
- Foundation for advanced reasoning: Mistral Small 3.1 enables further downstream customization of the model to create excellent reasoning models.
Mistral Small 3.1 can be used across various enterprise and consumer applications that require multimodal understanding, such as document verification, diagnostics, on-device image processing, visual inspection for quality checks, object detection in security systems, image-based customer support, and general-purpose assistance.
Mistral Small 3.1 is available to download on the Hugging Face website. It is also available via API on Mistral AI’s developer playground La Plateforme and on Google Cloud Vertex AI. It will be available on NVIDIA NIM and Microsoft Azure AI Foundry in the coming weeks.
Links
Official: Mistral-small-3-1
Huggingface: Mistral-Small-3.1-24B-Base-2503
Huggingface instruct: Mistral-Small-3.1-24B-Instruct-2503
More AI news
