Textbook of AI

Abstract. OpenAI's December 2024 announcement of o3. The flagship result was a step-change in performance on hard reasoning benchmarks given enough thinking budget, ARC-AGI, FrontierMath, GPQA all saw multi-fold improvements over o1. The decisive lesson was that the same underlying model could be made dramatically more capable simply by allowing it to spend more compute at the moment of answering. The release confirmed inference-time compute as a first-class scaling axis alongside training compute and parameter count, and reframed industry roadmaps around test-time scaling.