The 3B and 8B parameter Ministral-3-Instruct models include vision encoders to support multimodal understanding with multilingual instruction following across numerous languages. The 3B model targets ultra-lightweight edge environments fitting into 8GB VRAM, while the 8B model offers improved reasoning and memory-efficient inference within 12GB VRAM.
Both models support advanced features like structured function calling with JSON outputs under an open-source license, enabling resource-constrained and edge AI deployments with high performance. Customers can quickly deploy these models within AWS SageMaker for diverse AI workloads.
Source & further reading
Ministral-3-3B-Instruct-2512 and Ministral-3-8B-Instruct-2512 models now available on Amazon SageMaker JumpStart ↗Source published 2026-09-14. This report was prepared with AI assistance from the linked primary-source material. Claims and limitations should be checked against the original announcement or research.