Your AI SaaS business, ready by this week!
Try demoYour AI SaaS business, ready by this week!
Try demoA reasoning model is an LLM trained to think through a problem step by step before answering, which improves results on maths, code and planning at the cost of more time and tokens.
A reasoning model produces internal "thinking" tokens, working through the problem, before it writes the visible answer. Many let you set a reasoning effort or a thinking budget.
They are much stronger on multi-step problems: maths, debugging, analysis and planning. For simple questions they are slower and more expensive with little gain.
Thinking tokens are usually billed as output tokens, so one reasoning answer can cost many times a normal one. AI products need pricing that reflects that.
Reasoning models from providers such as OpenAI, Anthropic, Google and xAI are available like any other model in Aikeedo, metered in credits from their actual token use.
Discover more terms connected to this one
An LLM (large language model) is an AI model trained on very large amounts of text to predict the next token, which lets it write, summarise, translate, reason and hold a conversation.
Tokens are the small pieces of text an AI model reads and writes, roughly four characters or three quarters of an English word each, and the unit most AI providers bill by.
Chain of thought is when a model works through intermediate reasoning steps before giving its answer, either because it is prompted to or because it was trained to.