GPT-4o (“o” for “omni”) is OpenAI’s latest multimodal model, supporting both text and image inputs with text outputs. Delivers improved performance in non-English languages and visual understanding, while being faster and more cost-effective than previous models.