Models/
Mistral Releases Open-Weight Pixtral Vision-Language Model
Mistral AI has open-sourced Pixtral, a multimodal model capable of understanding images alongside text. The model supports document analysis, chart interpretation, and visual question answering at competitive quality levels. Released under Apache 2.0, Pixtral is available on Hugging Face and through Mistral's API platform. The release continues Mistral's strategy of building an open-weight ecosystem to compete with proprietary multimodal offerings from OpenAI and Google.