Beyond text and images
OpenAI's GPT-5 doesn't just "see" images — it reasons about them spatially, temporally, and contextually. Feed it a whiteboard sketch, and it'll turn it into working code with proper architecture.
The multimodal reasoning is genuinely impressive. It understands charts, diagrams, handwriting, screenshots, and video frames in context.
The real competition heats up
With Claude 4.5, Gemini 2.5, and now GPT-5 all shipping within weeks of each other, we're in the most competitive period in AI history. Builders win.