# Multi-image understanding

Phi-3.5-vision is a lightweight and advanced open-source multimodal model that supports a 128K context length and focuses on processing high-quality, inference-rich text and visual data.

Pixtral is a multimodal model based on the Mistral architecture that can handle image and text inputs and generate text outputs.

Pixtral-12B is a multimodal model compatible with the transformers library. It can handle image and text inputs and generate text outputs, suitable for image understanding and description tasks.