Mistral Small 3.1
Mistral Small 3.1, released in March 2025, is Mistral AI's compact but capable multimodal model featuring a 128K-token context window and integrated vision support. It is positioned as a highly efficient model that punches above its weight class, delivering strong instruction following and reasoning for a small model while also accepting image inputs for vision-language tasks. Mistral designed it to serve high-throughput production workloads where cost per token matters.
The 3.1 update brought meaningful improvements in instruction adherence, multilingual performance, and function calling reliability compared to earlier Mistral Small versions. Its vision capability allows developers to build applications that handle both text and image understanding within a single model, reducing architectural complexity. Available under an open license, it is suitable for self-hosted and cloud-served deployments alike.
Key Features
128K-token context window for long documents and extended sessions
Native vision support for image-and-text multimodal inputs
Reliable function calling and structured JSON output
Strong multilingual instruction following
Competitive performance-per-token ratio for high-throughput workloads
Open license supporting self-hosted and on-premises deployment
Ideal Use Cases
Cost-efficient production chatbots handling high request volumes
Document and image analysis in multimodal enterprise pipelines
Structured data extraction from mixed text-image content
Multilingual customer service automation
RAG systems requiring long-context document grounding
Example Prompts for Mistral Small 3.1
Technical Specifications
| Provider | Mistral |
| Category | Text |
| Modality | Text -> Text |
| Context Window | 128K tokens |
Frequently Asked Questions
Try Mistral Small 3.1 now
Start using Mistral Small 3.1 instantly — 100 free credits, no credit card required. Access 750+ AI models through one platform.