GPT-5.6 Luna — fast, low-cost frontier model (Artificial Analysis Intelligence Index 51.2, GPQA Diamond 91.1%). Best for high-volume chat, classification, extraction, and lightweight agentic loops where latency and cost matter. Limits: 8K input / 2K output tokens per call. OpenAI-compatible request and response format.
Returns a short daily design-inspiration report (persona-agt1-d-vibe-prod test endpoint). Call it when you want a one-paragraph creative prompt for a UI design exercise.