reasoning trace
In plain wordsThe written-out 'thinking' an AI reasoning model produces before its final answer — notes to itself full of attempts, checks and corrections, which users may or may not get to see.
The sequence of tokenstoken, n.: The small chunk of text an AI language model reads and writes — often a word, part of a word or a punctuation mark. AI usage, limits and prices are counted in tokens. a reasoning modelreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning. generates between receiving a prompt and starting its final answer, typically delimited by special tokensspecial token, n.: A reserved symbol in an AI model's vocabulary that is not ordinary text but a signal — 'start of message', 'end of answer', 'begin thinking' — used to structure what the model reads and writes. or returned as separate 'thinking' or 'reasoning' blocks. Traces contain problem decomposition, attempts, self-verification and backtracking, and may run to tens of thousands of tokens. Providers handle them differently: some return the full trace; others return a model-written summary, encrypted or redacted blocks that must be passed back unchanged in multi-turn or tool-using conversations, or only a token count for billing. Traces are used for knowledge distillationknowledge distillation, n.: Training a smaller or cheaper 'student' AI to copy the behavior of a larger 'teacher' AI, so the student keeps much of the teacher's skill while being faster to run. into smaller models, for debugging and for chain-of-thought monitoringchain-of-thought monitoring, n.: Having a program — often another AI — read the step-by-step reasoning an AI writes before it acts, to catch signs that it intends to cheat, deceive or cause harm.; how faithfully they reflect the model's actual computation is an open research question.
The reasoning trace showed it misread the units in step two and never went back.
Origin
An early prominent use is the ReActReAct, prop. n.: A method in which an AI alternates between thinking out loud about what to do next and taking an action, such as a web search, then uses what it observed to plan its next step. paper by Shunyu Yao and colleagues (October 2022), which interleaved reasoning traces with actions. The term became standard after OpenAI's o1 (September 2024), which hid its raw traces from users, and DeepSeek-R1DeepSeek-R1, prop. n.: A free, downloadable AI reasoning model released by the Chinese company DeepSeek in January 2025. It matched leading American models at math and coding for a fraction of the reported cost, shaking the industry. (January 2025), which exposed them; the s1 project (2025) fine-tuned a model on just 1,000 questions paired with traces.
Usage note
Often used interchangeably with chain of thoughtchain of thought, n.: Getting an AI to write out its reasoning step by step before giving an answer, the way a student shows their working. This often makes answers to math and logic problems more accurate.; some reserve 'reasoning trace' for the spontaneous output of trained reasoning models and 'chain of thought' for the prompted technique.
Related
- chain of thoughtchain of thought, n.: Getting an AI to write out its reasoning step by step before giving an answer, the way a student shows their working. This often makes answers to math and logic problems more accurate.
- reasoning modelreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning.
- extended thinkingextended thinking, n.: A setting in some AI models, notably Anthropic's Claude, that lets the model write out private step-by-step reasoning before its final answer, improving results on hard problems.
- chain-of-thought monitoringchain-of-thought monitoring, n.: Having a program — often another AI — read the step-by-step reasoning an AI writes before it acts, to catch signs that it intends to cheat, deceive or cause harm.
- chain-of-thought faithfulnesschain-of-thought faithfulness, n.: Whether the reasoning an AI writes out really reflects how it reached its answer, rather than being an after-the-fact story that leaves out what actually drove the decision.
- monitorabilitymonitorability, n.: How easy it is to watch an AI's reasoning and spot when it is planning something bad — a property that could be lost if future AIs stop thinking in readable language.
- latent reasoninglatent reasoning, n.: Getting an AI to do its step-by-step thinking inside its internal numbers rather than writing it out in words — potentially faster and more flexible, but much harder for people to inspect.
Sources
- OpenAI (2024). Learning to Reason with LLMs. OpenAI blog. openai.com/index/learning-to-reason-with-llms
- Muennighoff, N. et al. (2025). s1: Simple test-time scaling. arXiv:2501.19393. arxiv.org/abs/2501.19393
Mentioned in
chain of thoughtchain of thought, n.: Getting an AI to write out its reasoning step by step before giving an answer, the way a student shows their working. This often makes answers to math and logic problems more accurate., reasoning modelreasoning model, n.: An AI model trained to work through a problem step by step in hidden 'thinking' text before it answers. Letting it think longer generally makes it better at math, coding and planning., chain-of-thought monitoringchain-of-thought monitoring, n.: Having a program — often another AI — read the step-by-step reasoning an AI writes before it acts, to catch signs that it intends to cheat, deceive or cause harm., reasoningreasoning, n.: Working out conclusions step by step from facts, rules or evidence, as in solving a math problem or planning a trip. Whether today's AI truly reasons, or imitates reasoning, is debated., scratchpadscratchpad, n.: A space where an AI writes rough working — calculations, plans, notes — before or while producing its answer, much as a person jots on scrap paper., special tokenspecial token, n.: A reserved symbol in an AI model's vocabulary that is not ordinary text but a signal — 'start of message', 'end of answer', 'begin thinking' — used to structure what the model reads and writes., test-time computetest-time compute, n.: The computing effort an AI model spends while answering a question, as opposed to while being trained. Letting a model think longer or try many answers can make it markedly smarter., thinking budgetthinking budget, n.: A limit, set by whoever uses an AI model, on how much private step-by-step reasoning it may do before answering: more thinking usually means better answers but higher cost and slower replies., chain-of-thought faithfulnesschain-of-thought faithfulness, n.: Whether the reasoning an AI writes out really reflects how it reached its answer, rather than being an after-the-fact story that leaves out what actually drove the decision., extended thinkingextended thinking, n.: A setting in some AI models, notably Anthropic's Claude, that lets the model write out private step-by-step reasoning before its final answer, improving results on hard problems., interleaved thinkinginterleaved thinking, n.: A way for an AI agent to pause and reason again after every action it takes — reading each search result or test output before deciding what to do next — instead of only thinking once at the start., latent reasoninglatent reasoning, n.: Getting an AI to do its step-by-step thinking inside its internal numbers rather than writing it out in words — potentially faster and more flexible, but much harder for people to inspect., overthinkingoverthinking, n.: When an AI reasoning model spends far more thinking than a problem needs — pages of deliberation for a simple sum — wasting time and money and sometimes talking itself out of a right answer.