Reasoning model APIs, distinct from standard large language models (LLMs), are designed to perform complex logical tasks by generating step-by-step chains of thought. These models excel in applications such as planning, coding, math proofs, and research, offering more reliable results but requiring longer context windows and incurring higher computational costs. Notable models include OpenAI's O-series, Google DeepMind's Gemini 2.5 Pro, and Anthropic's Claude Opus 4, each offering unique strengths in context length, cost, and performance. Open-source options like Llama 4 Scout and DeepSeek R1 allow for self-hosting, providing privacy and customization. Despite higher costs, reasoning models are gaining traction due to their ability to handle multimodal inputs, such as text and images, and their integration with advanced tools and safety measures. Clarifai supports these models through compute orchestration and evaluation tools, assisting users in deploying reasoning capabilities effectively while managing costs.