OpenAI released two new artificial intelligence models Wednesday, marking a leap in multimodal reasoning capabilities, according to CNBC. The new model can analyze and respond to user-uploaded sketches, diagrams, and whiteboard images, even at low quality. It integrates visual information into its reasoning process. OpenAI said the model is tuned for math, science, coding, and image interpretation. Another model is faster, lower cost, and built for efficiency. OpenAI said they passed the company’s most rigorous safety testing yet, though recent policy changes eliminated some reporting requirements.