Pixtral Large is a cutting-edge multimodal AI model introduced by Mistral AI, built upon Mistral Large 2. It features advanced image understanding capabilities, enabling comprehension of documents, charts, and natural images while retaining Mistral Large 2's leadership in text understanding. The model has demonstrated exceptional performance in multimodal benchmarks, surpassing other models in tests such as MathVista, ChartQA, and DocVQA. It has also shown competitiveness in the MM-MT-Bench tests, outperforming various models, including Claude-3.5 Sonnet. The model is available under the Mistral Research License (MRL) for research and educational purposes and the Mistral Commercial License for commercial use.