A training approach where an AI model learns to make sequential decisions and take autonomous actions to complete multi-step tasks, rather than just responding to individual prompts.
Code generation, debugging, explanation, and refactoring
Multi-step reasoning, logic puzzles, mathematical problem-solving